Pith. sign in

REVIEW 4 major objections 6 minor 146 references

ChronoLens: Measuring Language Change Across Time, Languages, and Linguistic Levels

T0 review · 4 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read A single feature-aligned space measures language change across five languages and four linguistic levels at once, revealing that the levels shift by comparable amounts within a language while languages differ in timing, magnitude, and direc

desk verdict A genuinely useful framework for comparing diachronic change across languages and levels, but the pragmatics results contradict the paper's own feature-count criterion and the validation is partly circular. read the letter →

arxiv 2608.03507 v1 pith:5G5PNTD3 submitted 2026-08-04 cs.CL cs.AI

classification cs.CLcs.AI
keywords languagechangediachroniclinguisticscrosscoderssparseautoencodersmultilingualmodelsparliamentarycorporacross-linguisticcomparisonlinguisticlevels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

ChronoLens asks a question earlier computational work could not answer: do morphology, syntax, semantics, and pragmatics change together over time, or do they follow separate historical paths? The paper's claim is that they can be measured together — by encoding sentences from five parliamentary traditions with frozen multilingual language models and projecting them into one feature-aligned sparse space learned by crosscoders — and that when measured together, the four levels change by roughly comparable amounts within a language, while languages differ sharply in when, how far, and in which direction they change. The framework's representations agree with direct linguistic statistics far better than dense embeddings or a pooled sparse autoencoder (Spearman ρ = 0.72 versus 0.29 and 0.28), which the paper reads as evidence that its features track real linguistic variation rather than reconstruction artifacts. If the magnitude–direction separation is real, then single-language, single-level studies have been under-reporting the structure of historical change, because similar magnitudes can conceal very different trajectories.

What carries the argument

The load-bearing object is the crosscoder: a sparse autoencoder with one shared feature index across all conditions — languages in the cross-lingual setting, or historical periods within a language — but separate encoders and decoders per condition. A single sparse feature vector f is inferred jointly from a tuple of matched sentences (ReLU plus BatchTopK), so feature j carries the same identity in every condition, while each condition's decoder lets that feature contribute with different strength there. This shared-index design is what makes cross-lingual and cross-temporal comparison possible, since independently trained sparse dictionaries have no guaranteed feature correspondence. Lingui

What would settle it

Train the pipeline on a condition-shuffled null — tuples whose decade or language identities are randomly exchanged, a control the paper already runs at feature level — and apply the full magnitude-and-direction analysis to it. If the null reproduces the ρ ≈ 0.72 linguistic agreement or the German-versus-Turkish magnitude split, the signal is a method artifact, not a fact about time. A complementary check: apply the pipeline to a decades-spanning corpus of stable, repeatedly reprinted texts (legal or liturgical boilerplate) with a similar OCR profile; if it reports substantial change, the traj

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a crosscoder — a sparse autoencoder whose feature indices are shared across conditions but which gives each language or period its own decoder — yields representations whose decade-to-decade movement tracks independently measured linguistic statistics more closely than dense embeddings or a pooled sparse autoencoder do (mean Spearman ρ = 0.72 versus 0.29 and 0.28; linguistic specificity 0.89 versus 0.74 and 0.73). Applied to 44.98 million parliamentary documents and about 17.2 billion tokens across English, German, Italian, Polish, and Turkish (1803–2026), this representation shows that within a language the four linguistic levels change by com

Load-bearing premise

The measured displacements are read as language change rather than corpus change: the five parliamentary traditions have unequal historical coverage, different source mixes, and OCR noise concentrated in the earliest material, so if length- and topic-matching do not fully control those differences, the magnitude and direction results would describe corpus artifacts, not language change.

Editorial extensions

If this is right

  • If the ρ = 0.72 agreement is real, feature-aligned sparse representations are the appropriate substrate for diachronic comparison, and the pooled autoencoder's failure shows that a shared sparse dictionary alone, without condition-specific decoders, is not enough.
  • Within each language, morphology, syntax, semantics, and pragmatics change by comparable amounts, so single-level histories (the common practice) systematically miss that the other levels are moving just as much.
  • Magnitude does not determine direction: Italian and Turkish share net magnitude 0.53 but move in different directions, while German (0.89) and Turkish (0.53) move different distances in similar directions, so cross-linguistic comparison must report both quantities.
  • Pragmatics is the most shared dimension of change across the five languages (mean pairwise cosine 0.92, with personal deixis rising in all five), while syntax shows no common direction (0.00), suggesting shared institutional or cultural pressures shape parliamentary language beyond genealogical relatedness.
  • Rankings of which language changed most depend on the historical window: German is largest over its full record, Turkish over the common 1950–2020 interval, so claims about relative change are interval-relative.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's own threshold-sensitivity analysis reports that features assigned to pragmatics are scarce (0.4–0.6% of assigned features at every threshold setting), so the claim that pragmatics changes in step with the other levels may rest on a thinner feature base; a dedicated pragmatic feature bank would test whether the high cross-language pragmatics agreement (cosine 0.92) survives a denser inv
  • The hard one-level-per-feature assignment is a measurement choice, not an established fact about language: genuinely cross-level features, which the paper acknowledges exist, are forced into a dominant level, so phenomena like grammaticalization that plausibly run through morphology, syntax, and semantics at once may be under-counted in the reported trajectories.
  • The headline direction metric compares net displacement between first and last decades, so pairs like German and Turkish could reach similar bearings through different intermediate paths; the paper's secondary stepwise-alignment measure is the stricter test of whether two languages truly follow shared trajectories.
  • Because directionally similar pairs (English–Italian, German–Turkish) are not genealogically close, the framework offers a quantitative handle on contact and institutional-convergence hypotheses — for instance, whether shared European political discourse is steering parliamentary registers toward a common pragmatic style, which the pragmatics results make plausible but do not causally establish.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper introduces ChronoLens, a framework combining frozen multilingual language models with feature-aligned crosscoders to measure historical language change across five parliamentary traditions (English, German, Italian, Polish, Turkish) and four linguistic levels (morphology, syntax, semantics, pragmatics). The corpus spans 1803–2026 (44.98M documents, ~17.2B tokens). The core claims are: (i) the crosscoder's sparse representations agree more strongly with linguistic statistics than dense embeddings or a pooled SAE (ρ=0.72 vs 0.29/0.28; Table 2); (ii) within a language, the four linguistic levels change by comparable magnitudes, while languages differ in timing, magnitude, and direction (Figures 2–4); and (iii) magnitude and direction of language change are separable properties. The framework trains crosscoders without linguistic supervision, then assigns features to linguistic levels via post-hoc probe ablations, and computes magnitude/direction metrics in the shared feature space.

Significance. If the claims hold, ChronoLens offers a genuinely unifying representation for cross-linguistic, cross-level diachronic comparison, addressing a recognized gap in computational historical linguistics. The paper's strengths include its very large harmonized corpus, the use of four diverse multilingual backbones, explicit condition-shuffled nulls and permuted-label controls, and a transparent threshold-sensitivity analysis. The empirical finding that magnitude and direction are separable is important and falsifiable. However, the central validation and the four-level empirical claim are currently undermined by two problems: the 'independent' linguistic indicators are derived from the same annotations used to train the assignment probes and select layers, and the paper's own Appendix C.4 states that pragmatics trajectories never meet the minimum feature-count criterion, directly contradicting the abstract's four-level claim. The contribution is potentially significant, but the evidence as presented does not yet support the headline conclusions.

major comments (4)
  1. [Appendix C.4; §5.2, Fig. 3; Abstract] The threshold-sensitivity analysis reports that pragmatics features are only 0.4–0.6% of assigned features at every threshold setting and 'never reaches that count', so 'the filtered trajectories cover morphology, syntax and semantics only.' This directly contradicts the abstract's statement that morphology, syntax, semantics, and pragmatics 'generally change by comparable amounts,' and Figure 3's pragmatics magnitudes (e.g., 0.29–0.50) are computed from cells that, by C.6's own rule, fall below the minimum of eight available features. Either the figure reports unstable low-coverage cells, or the appendix refers to a different analysis; the manuscript must reconcile this and either remove pragmatics from the four-level claim or demonstrate stable trajectories with adequate feature coverage.
  2. [§4.2, Table 2; Appendix D; §4.1/C.2] The headline validation metric 'linguistic agreement' is described as correlating representational displacement with 'direct changes in independently measured linguistic indicators.' However, Appendix D states that the observable measures are 'derived from the same sentence-level annotations used for feature attribution,' and §4.1 and Appendix C.2 show that both layer selection and feature-to-level assignment use probes trained on those same annotations. Features are therefore selected because they predict the labels against which agreement is measured, so ρ=0.72 is partly by construction. The comparison with embeddings/SAE remains informative, but the claim of independent linguistic grounding requires a validation source held out from probe training and layer selection, or an explicit demonstration that the agreement survives such a split.
  3. [§3, §4.1; Limitations] The empirical trajectories are interpreted as language change, but corpus comparability is not established. Coverage differs markedly by language (English begins 1803, Turkish 1950), German additionally includes German Digital Library newspapers, all languages contain manifesto records in different proportions, and OCR noise is concentrated in the earliest material. Length- and policy-frame matching controls for some sentence-level confounds, but not for source composition, register, or OCR degradation. The Limitations admit that 'residual OCR errors... may still resemble linguistic change despite our filtering.' Since the central findings concern magnitudes and directions over time, the paper should include per-source or per-register stability analyses (e.g., parliamentary speeches only, common periods) and show that the trajectories are not driven by OCR or source shifts.
  4. [Appendix C.5, Table 9; §5.2, Fig. 3] Even setting aside the feature-count issue, the pragmatics attribution rests on weak probes: speech-act selectivity ranges 0.05–0.21 and the majority class covers 0.84–0.96 of held-out examples, and the paper itself cautions that lower-selectivity tasks should be interpreted with caution. Reporting pragmatics magnitudes as comparable to other levels (Figure 3) is not supported by the released probe diagnostics. At minimum, the pragmatics rows should be flagged as low-confidence and the cross-level 'comparable amounts' claim should be restricted to the levels with reliable attribution.
minor comments (6)
  1. [§5.2 vs Appendix D] The main text says 'We examine 23 measures grouped under morphology, syntax, semantics, and pragmatics' and '37 of the 90 fitted trends remain significant.' Appendix D reports 18 observable measures, and 18×5=90; 23 measures would give 115 tests. The mismatch should be corrected.
  2. [Appendix C.6] The variable name 'parend' appears to be a typo; the surrounding text and Eq. (15) define a cosine between displacement vectors. Please rename for clarity.
  3. [Table 2] The row 'Linguistic agreement, ρ↑' does not state whether ρ is Spearman or Pearson. The text calls it 'mean Spearman correlation' in §4.2 but Table 2 is ambiguous. Please state the estimator in the table caption.
  4. [Fig. 3; Appendix C.6] Cells with fewer than the minimum number of features are retained but marked as low coverage according to C.6, yet Figure 3 does not mark any low-coverage cells. Adding a visual indicator for low-coverage cells would make the reliability of the pragmatics rows transparent.
  5. [Appendix C.4] The text refers to 'the pre-registered target × decade convergence contrast' but no pre-registration identifier or repository link is provided. Please add the registration reference or remove the term.
  6. [§5.2] Typo: 'ross-language agreement' should be 'cross-language agreement.'

Circularity Check

1 steps flagged · score 6.0 of 10

Linguistic-agreement validation is partly circular: probe-selected features are correlated with the same annotations used to select them; pragmatics trajectories are additionally unsupported by the paper's own feature-count criterion.

  1. fitted input called prediction [Section 4.2 (Linguistic agreement) / Appendix D / Eqs. (3) and (13)]
    "Linguistic agreement is the mean Spearman correlation between decade-to-decade representational displacement and direct changes in independently measured linguistic indicators for morphology, syntax, semantics, and pragmatics … To interpret these aggregate changes, we also examine 18 observable linguistic measures derived from the same sentence-level annotations used for feature attribution."

    Features entering the trajectory uλ (Eq. 3) are assigned to level λ by Eq. (13), which selects features whose ablation most decreases correct-label probability under probes trained on the very annotations that Appendix D later calls the 'observable linguistic measures'. The 'independently measured linguistic indicators' are therefore the same labels that drove feature attribution and layer selection. The reported ρ=0.72 measures correlation between activation trajectories of label-informative features and changes in the labels that made them informative, so the agreement is partly a selection artifact rather than an external confirmation. Running the embedding and SAE baselines through the same selection/validation loop does not remove this; it only makes the comparison less unfair, not in

full rationale

The core architecture is not circular: the crosscoder is trained without linguistic labels, the shared-feature/condition-decoder construction is a genuine modeling choice, and the magnitude/direction decomposition is a mathematical definition applied to learned representations. No load-bearing self-citation chain or imported uniqueness theorem appears; self-references (Eger & Mehler, Walter et al., Reinig et al., Chen et al.) are used for specific tasks or prior patterns, not to forbid alternatives. The main circularity is the headline validation metric: feature attribution (Eq. 13) selects features by ablation effects on probes trained on the same annotations that the linguistic-agreement metric correlates with, and Appendix D explicitly derives the 'observable' measures from those same annotations. Thus the central claim 'crosscoder agrees more strongly with linguistic statistics' is partially forced by the selection procedure. Separately, Appendix C.4 states that pragmatics never reaches the minimum-feature count at any threshold ('so the filtered trajectories cover morphology, syntax and semantics only'), while Figure 3 and the abstract report pragmatics magnitudes; this is an internal support gap rather than a definitional circularity, but it weakens the four-level empirical claim. On balance, the representational-validity claim is partially circular (score 6), while the direction/magnitude findings and cross-lingual comparisons retain independent empirical content.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claims rest on domain assumptions about corpus comparability, parser accuracy on historical text, and cross-lingual feature alignment, plus a set of hand-chosen hyperparameters. No new theoretical entities (particles, forces, dimensions) are introduced; the 'linguistic levels' are operational constructs from standard categories.

free parameters (8)
  • Crosscoder dictionary expansion factor (2x hidden dimension) = 2
    Sets the number of shared sparse features; no sensitivity analysis for this choice is reported.
  • BatchTopK active fraction = 0.10
    Controls sparsity; no sensitivity analysis for this value is reported.
  • Attribution absolute threshold tau_abs and relative separation tau_rel = 0.01 and 1.2
    Used for linguistic level assignment; sensitivity analysis shows main conclusions are stable across settings, but this is still a hand-chosen cutoff.
  • Backbone layer selection by probe selectivity = L9/L16/L20/L10 per backbone
    Layer is chosen to maximize probe selectivity on the same probe tasks used for attribution and validation, a form of model selection on the evaluation labels (Appendix C.2).
  • Decoder norm ratio threshold for shared/specific features = 4
    Threshold separating shared from condition-specific features; sensitivity checked at 2, 3, 6, 8 but is still a free choice.
  • Per-decade sentence cap and minimum distinct sentences = 400 per concept/language/decade; 250
    Temporal balancing choices affecting which decades are usable as crosscoder conditions.
  • Temporal selectivity AUC and split-half robustness thresholds = AUC >= 0.55; Spearman >= 0.4
    Diagnostic gates for feature reliability, not part of the main trajectory filter but still arbitrary cutoffs.
  • Alignment parallel/anti-parallel thresholds = +/-0.20
    Decision thresholds for labeling language-pair direction alignment in Appendix D.
assumptions (5)
  • domain assumption A single frozen multilingual language model places all five languages in a sufficiently shared representational space so that crosscoder feature indices are comparable across languages and periods.
    Load-bearing for all cross-lingual and diachronic comparisons; invoked in Section 4.1 and Appendix C.2 with no direct validation of cross-lingual feature identity.
  • domain assumption Universal Dependencies parsers and multilingual lexical resources yield accurate linguistic annotations on 1803-2026 parliamentary texts across five languages.
    These annotations generate the probe labels and the 18 observable measures; parser accuracy on historical text is not validated and could create artificial change signals.
  • domain assumption Parliamentary registers across five countries and 200+ years are comparable after matching on token length and policy frame.
    Central to interpreting trajectories as language change rather than register or corpus-composition change; the paper acknowledges residual OCR and source-mix issues in the Limitations.
  • domain assumption The crosscoder architecture (condition-specific encoders, summed representation, BatchTopK, condition-specific decoders) yields stable and genuinely aligned features rather than an artifact of the pooling operation.
    The paper provides a condition-shuffled null and latent-scaling checks, but the architectural assumption remains foundational; Section 4.1 and Appendix C.3.
  • standard math Standard mathematical background for linear algebra, sparse coding, simulation-based inference (Spearman correlation, bootstrap, mixed-effects models).
    Unproved background results relied on throughout the evaluation sections.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ChronoLens: Measuring Language Change Across Time, Languages, and Linguistic Levels." pith.science (2026). https://pith.science/paper/5G5PNTD3

@misc{pith2026260803507,
  author       = {Pith},
  title        = {Pith review of: ChronoLens: Measuring Language Change Across Time, Languages, and Linguistic Levels},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5G5PNTD3}},
  note         = {Machine review of arXiv:2608.03507}
}
abstract

Historical language change affects morphology, syntax, semantics, and pragmatics, yet computational studies typically examine these levels with incompatible representations and therefore cannot determine whether they evolve together across languages. We address this problem by asking how the magnitude and direction of change vary across linguistic levels, languages, and historical periods within a single analytical space. We introduce ChronoLens, a framework that combines frozen multilingual language models, feature-aligned crosscoders, and post-hoc linguistic interventions, and apply it to 44.98 million documents and approximately 17.2 billion tokens from five parliamentary traditions spanning 1803--2026. The resulting sparse representations agree substantially more strongly with linguistic statistics than dense embeddings or a pooled sparse autoencoder ($\rho=0.72$ versus $0.29$ and $0.28$), and reveal that morphology, syntax, semantics, and pragmatics generally change by comparable amounts within a language, while languages differ markedly in when, how far, and in which direction they change. These findings show that historical language change is a structured, multidimensional process: similar magnitudes can conceal different trajectories, and meaningful cross-linguistic comparison requires measuring both distance and direction.

Figures

Figures reproduced from arXiv: 2608.03507 by the authors.

Figure 1
Figure 1. Overview of crosscoder training and post-hoc linguistic attribution. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. Magnitude of historical change between 1950 [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 2
Figure 2. Magnitude of historical change over each [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Magnitude and direction of historical change [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Tokens per decade (Calculated using Llama-3 [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Observable linguistic changes between 1950 and 2020. The left panel reports the fitted change in 18 [PITH_FULL_IMAGE:figures/full_fig_p019_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

146 extracted references · 54 canonical work pages

  1. [1]

    Aho and Jeffrey D

    Alfred V. Aho and Jeffrey D. Ullman , title =. 1972

  2. [2]

    Publications Manual , year = "1983", publisher =

  3. [3]

    Chandra and Dexter C

    Ashok K. Chandra and Dexter C. Kozen and Larry J. Stockmeyer , year = "1981", title =. doi:10.1145/322234.322243

  4. [4]

    Scalable training of

    Andrew, Galen and Gao, Jianfeng , booktitle=. Scalable training of

  5. [5]

    Dan Gusfield , title =. 1997

  6. [6]

    Tetreault , title =

    Mohammad Sadegh Rasooli and Joel R. Tetreault , title =. Computing Research Repository , volume =. 2015 , url =

  7. [7]

    A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =

    Ando, Rie Kubota and Zhang, Tong , Issn =. A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =. Journal of Machine Learning Research , Month = dec, Numpages =

  8. [8]

    Optimizing Statistical Machine Translation for Text Simplification , journal =

    Wei Xu and Courtney Napoles and Ellie Pavlick and Quanze Chen and Chris Callison. Optimizing Statistical Machine Translation for Text Simplification , journal =. 2016 , doi =

Show all 146 references
  1. [9]

    Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics , pages =

    Kishore Papineni and Salim Roukos and Todd Ward and Wei. Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics , pages =. 2002 , publisher =. doi:10.3115/1073083.1073135 , url =

  2. [10]

    Proceedings of the Third Conference on Machine Translation: Research Papers (WMT) , pages =

    Matt Post , title =. Proceedings of the Third Conference on Machine Translation: Research Papers (WMT) , pages =. 2018 , publisher =. doi:10.18653/v1/W18-6319 , url =

  3. [11]

    Manning , title =

    Peng Qi and Yuhao Zhang and Yuhui Zhang and Jason Bolton and Christopher D. Manning , title =. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations , pages =. 2020 , publisher =. doi:10.18653/v1/2020.acl-demos.14 , url =

  4. [12]

    Universal Dependencies v2: An Evergrowing Multilingual Treebank Collection , booktitle =

    Joakim Nivre and Marie. Universal Dependencies v2: An Evergrowing Multilingual Treebank Collection , booktitle =. 2020 , publisher =

  5. [13]

    Toma. The. Language Resources and Evaluation , volume =. 2023 , doi =

  6. [14]

    1985 , doi =

    Jeffrey Scott Vitter , title =. 1985 , doi =

  7. [15]

    1991 , doi =

    Jianhua Lin , title =. 1991 , doi =

  8. [16]

    The Twelfth International Conference on Learning Representations (ICLR) , year =

    Robert Huben and Hoagy Cunningham and Logan Riggs Smith and Aidan Ewart and Lee Sharkey , title =. The Twelfth International Conference on Learning Representations (ICLR) , year =

  9. [17]

    Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing (EMNLP-IJCNLP) , pages =

    Nils Reimers and Iryna Gurevych , title =. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing (EMNLP-IJCNLP) , pages =. 2019 , publisher =. doi:10.18653/v1/D19-1410 , url =

  10. [18]

    Hamilton and Jure Leskovec and Dan Jurafsky , title =

    William L. Hamilton and Jure Leskovec and Dan Jurafsky , title =. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages =. 2016 , publisher =. doi:10.18653/v1/P16-1141 , url =

  11. [19]

    Gonzalez and Hao Zhang and Ion Stoica , title =

    Woosuk Kwon and Zhuohan Li and Siyuan Zhuang and Ying Sheng and Lianmin Zheng and Cody Hao Yu and Joseph E. Gonzalez and Hao Zhang and Ion Stoica , title =. Proceedings of the 29th Symposium on Operating Systems Principles (SOSP) , pages =. 2023 , publisher =. doi:10.1145/3600...

  12. [20]

    Lianmin Zheng and Wei. Judging. Advances in Neural Information Processing Systems 36 (NeurIPS) , year =

  13. [21]

    CoRR , volume =

    The. CoRR , volume =. 2024 , doi =

  14. [22]

    2025 , doi =

    Qwen3 Technical Report , journal =. 2025 , doi =

  15. [23]

    Albert Q. Jiang and Alexandre Sablayrolles and Arthur Mensch and Chris Bamford and Devendra Singh Chaplot and Diego de Las Casas and Florian Bressand and Gianna Lengyel and Guillaume Lample and Lucile Saulnier and L. Mistral 7B , journal =. 2023 , doi =

  16. [24]

    CoRR , volume =

    Pedro Henrique Martins and Patrick Fernandes and Jo. CoRR , volume =. 2024 , doi =

  17. [25]

    Journal of Cognitive Science , volume =

    Haitao Liu , title =. Journal of Cognitive Science , volume =

  18. [26]

    Journal of Applied Psychology , volume =

    Rudolf Flesch , title =. Journal of Applied Psychology , volume =. 1948 , doi =

  19. [27]

    Scuola e Citt

    Pietro Lucisano and Maria Emanuela Piemontese , title =. Scuola e Citt

  20. [28]

    Richard Bamberger and Erich Vanecek , title =

  21. [29]

    Educational and Psychological Measurement , volume =

    Jacob Cohen , title =. Educational and Psychological Measurement , volume =. 1960 , doi =

  22. [30]

    Transformer Circuits Thread , year =

    Trenton Bricken and Adly Templeton and Joshua Batson and Brian Chen and Adam Jermyn and others , title =. Transformer Circuits Thread , year =

  23. [31]

    Transformer Circuits Thread , year =

    Adly Templeton and Tom Conerly and Jonathan Marcus and Jack Lindsey and Trenton Bricken and others , title =. Transformer Circuits Thread , year =

  24. [32]

    A Systematic Comparison of Contextualized Word Embeddings for Lexical Semantic Change , author =. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) , month = ju...

  25. [33]

    Proceedings of the 5th Workshop on Computational Approaches to Historical Language Change , month = aug, year =

    Towards a Complete Solution to Lexical Semantic Change: an Extension to Multiple Time Periods and Diachronic Word Sense Induction , author =. Proceedings of the 5th Workshop on Computational Approaches to Historical Language Change , month = aug, year =

  26. [34]

    Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , month = aug, year =

    Analyzing Semantic Change through Lexical Replacements , author =. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , month = aug, year =

  27. [35]

    Proceedings of the 31st International Conference on Computational Linguistics , month = jan, year =

    Analyzing Continuous Semantic Shifts with Diachronic Word Similarity Matrices , author =. Proceedings of the 31st International Conference on Computational Linguistics , month = jan, year =

  28. [36]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , month = jul, year =

    Quantifying Lexical Semantic Shift via Unbalanced Optimal Transport , author =. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , month = jul, year =

  29. [37]

    Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025) , month = may, year =

    Lexical Semantic Change Annotation with Large Language Models , author =. Proceedings of the 9th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature (LaTeCH-CLfL 2025) , month = may, year =

  30. [38]

    2025 , address =

    Baes, Naomi and Merx, Raphael and Haslam, Nick and Vylomova, Ekaterina and Dubossarsky, Haim , booktitle =. 2025 , address =

  31. [39]

    Findings of the Association for Computational Linguistics: ACL 2025 , month = jul, year =

    Narrative Media Framing in Political Discourse , author =. Findings of the Association for Computational Linguistics: ACL 2025 , month = jul, year =

  32. [40]

    and Mahdi, Adam , year =

    Ghafouri, Vahid and McNeil, Robert and Yankov, Teodor and Sumption, Madeleine and Rocher, Luc and Hale, Scott A. and Mahdi, Adam , year =. Framing Migration: A Computational Analysis of. 2509.14197 , archivePrefix =

  33. [41]

    Metaphorical Framing of Refugees, Asylum Seekers and Immigrants in

    Wang, Yunxiao , booktitle =. Metaphorical Framing of Refugees, Asylum Seekers and Immigrants in. 2024 , address =

  34. [42]

    Proceedings of the First Workshop on Reference, Framing, and Perspective @ LREC-COLING 2024 , month = may, year =

    Comparing News Framing of Migration Crises using Zero-Shot Classification , author =. Proceedings of the First Workshop on Reference, Framing, and Perspective @ LREC-COLING 2024 , month = may, year =

  35. [43]

    Fine-Grained Detection of Solidarity for Women and Migrants in 155 Years of

    Kostikova, Aida and Paassen, Benjamin and Beese, Dominik and P. Fine-Grained Detection of Solidarity for Women and Migrants in 155 Years of. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing , month = nov, year =

  36. [44]

    Proceedings of the IV Workshop on Creating, Analysing, and Increasing Accessibility of Parliamentary Corpora (ParlaCLARIN) @ LREC-COLING 2024 , month = may, year =

    Resources and Methods for Analysing Political Rhetoric and Framing in Parliamentary Debates , author =. Proceedings of the IV Workshop on Creating, Analysing, and Increasing Accessibility of Parliamentary Corpora (ParlaCLARIN) @ LREC-COLING 2024 , month = may, year =

  37. [45]

    Syntactic Language Change in

    Chen, Yanran and Zhao, Wei and Breitbarth, Anne and Stoeckel, Manuel and Mehler, Alexander and Eger, Steffen , year =. Syntactic Language Change in. 2402.11549 , archivePrefix =

  38. [46]

    Tracing Syntactic Complexity: Exploring the Evolution of Average Dependency Length Across Three Centuries of Scientific

    Krielke, Marie-Pauline and Alves, Diego and Talamo, Luigi , booktitle =. Tracing Syntactic Complexity: Exploring the Evolution of Average Dependency Length Across Three Centuries of Scientific. 2025 , address =

  39. [47]

    Proceedings of the 6th Workshop on Research in Computational Linguistic Typology and Multilingual NLP , month = mar, year =

    Syntactic Dependency Length Shaped by Strategic Memory Allocation , author =. Proceedings of the 6th Workshop on Research in Computational Linguistic Typology and Multilingual NLP , month = mar, year =

  40. [48]

    Humanities and Social Sciences Communications , volume =

    Dependency Distance Minimization: A Diachronic Exploration of the Effects of Sentence Length and Dependency Types , author =. Humanities and Social Sciences Communications , volume =. 2022 , url =

  41. [49]

    Proceedings of the 37th Pacific Asia Conference on Language, Information and Computation , month = dec, year =

    The Cross-linguistic Variations in Dependency Distance Minimization and its Potential Explanations , author =. Proceedings of the 37th Pacific Asia Conference on Language, Information and Computation , month = dec, year =

  42. [50]

    Languages in Contrast , volume =

    Cross-linguistic Dependency Length Minimization in Scientific Language , author =. Languages in Contrast , volume =. 2024 , url =

  43. [51]

    2024 , address =

    de Jong, Asher and Kuzman, Taja and Larooij, Maik and Marx, Maarten , booktitle =. 2024 , address =

  44. [52]

    2024 , eprint =

    Multilingual Power and Ideology Identification in the Parliament: A Reference Dataset and Simple Baselines , author =. 2024 , eprint =

  45. [53]

    Erjavec, Toma. The. Language Resources and Evaluation , volume =. 2022 , url =

  46. [54]

    Proceedings of the IV Workshop on Creating, Analysing, and Increasing Accessibility of Parliamentary Corpora (ParlaCLARIN) @ LREC-COLING 2024 , month = may, year =

    Parliamentary Discourse Research in Political Science: Literature Review , author =. Proceedings of the IV Workshop on Creating, Analysing, and Increasing Accessibility of Parliamentary Corpora (ParlaCLARIN) @ LREC-COLING 2024 , month = may, year =

  47. [55]

    Compiling and Exploring a

    Aires, Jos. Compiling and Exploring a. Proceedings of the IV Workshop on Creating, Analysing, and Increasing Accessibility of Parliamentary Corpora (ParlaCLARIN) @ LREC-COLING 2024 , month = may, year =

  48. [56]

    Statistical Semantic Change Detection via Usage Similarities

    Aida, Taichi and Mochihashi, Daichi and Takamura, Hiroya and Ogiso, Toshinobu and Komachi, Mamoru. Statistical Semantic Change Detection via Usage Similarities. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). ...

  49. [57]

    DHPLT : large-scale multilingual diachronic corpora and word representations for semantic change modelling

    Fedorova, Mariia and Kutuzov, Andrey and Umarova, Khonzoda. DHPLT : large-scale multilingual diachronic corpora and word representations for semantic change modelling. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ...

  50. [58]

    Transparent Semantic Change Detection with Dependency-Based Profiles

    Tat, Bach Phan and Heylen, Kris and Geeraerts, Dirk and De Pascale, Stefano and Speelman, Dirk. Transparent Semantic Change Detection with Dependency-Based Profiles. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' ...

  51. [59]

    Semantic Change Characterization with LLM s using Rhetorics

    de S \'a , J \'a der Martins Camboim and Lee, Jooyoung and Da Silveira, Marcos and Pruski, Cedric. Semantic Change Characterization with LLM s using Rhetorics. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). 2...

  52. [60]

    Rethinking Metrics for Lexical Semantic Change Detection

    Goworek, Roksana and Dubossarsky, Haim. Rethinking Metrics for Lexical Semantic Change Detection. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). 2026. doi:10.18653/v1/2026.lchange-1.13

  53. [61]

    Pretraining Language Models for Diachronic Linguistic Change Discovery

    Fittschen, Elisabeth and Li, Sabrina Xin and Lippincott, Tom and Choshen, Leshem and Messner, Craig. Pretraining Language Models for Diachronic Linguistic Change Discovery. Findings of the Association for Computational Linguistics: EACL 2026. 2026. doi:10.18653/v1/2026.finding...

  54. [62]

    Threshold-Calibrated Word Sense Disambiguation: Semantic Broadening Without Sense Redistribution in Schizophrenia

    Baes, Naomi and Haslam, Nick. Threshold-Calibrated Word Sense Disambiguation: Semantic Broadening Without Sense Redistribution in Schizophrenia. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). 2026. doi:10.186...

  55. [63]

    Using BERT to Explore Lexical Semantic Change of Prepositions

    Radchankava, Liudmila and Konovalov, Vasily. Using BERT to Explore Lexical Semantic Change of Prepositions. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). 2026. doi:10.18653/v1/2026.lchange-1.10

  56. [64]

    and Dinu, Anca Daniela and Georgescu, Simona

    Uban, Ana Sabina and Dinu, Liviu P. and Dinu, Anca Daniela and Georgescu, Simona. Cross-lingual Lexical Semantic Change in R omance Languages. The Proceedings for the 6th International Workshop on Computational Approaches to Language Change ( LC hange ' 26). 2026. doi:10.18653...

  57. [65]

    PLOS One , volume =

    Syntactic language change in English and German: Metrics, parsers, and convergences , author =. PLOS One , volume =. 2026 , url =

  58. [66]

    Semantic Change and Emerging Tropes In a Large Corpus of N ew H igh G erman Poetry

    Haider, Thomas and Eger, Steffen. Semantic Change and Emerging Tropes In a Large Corpus of N ew H igh G erman Poetry. Proceedings of the 1st International Workshop on Computational Approaches to Historical Language Change. 2019. doi:10.18653/v1/W19-4727

  59. [67]

    On the Linearity of Semantic Change: Investigating Meaning Variation via Dynamic Graph Models

    Eger, Steffen and Mehler, Alexander. On the Linearity of Semantic Change: Investigating Meaning Variation via Dynamic Graph Models. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2016. doi:10.18653/v1/P16-2009

  60. [68]

    Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models

    Shi, Dan and Han, Zhuowen and Ostermann, Simon and Jin, Renren and Van Genabith, Josef and Xiong, Deyi. Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models. Proceedings of the 64th Annual Meeting of the A ssoc...

  61. [69]

    Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders

    Deng, Boyi and Wan, Yu and Yang, Baosong and Zhang, Yidan and Feng, Fuli. Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 202...

  62. [70]

    arXiv preprint arXiv:2507.11230 , year=

    Sparse Autoencoders Can Capture Language-Specific Concepts Across Diverse Languages , author=. arXiv preprint arXiv:2507.11230 , year=

  63. [71]

    Deciphering Cultural Representations in Large Language Models via Sparse Autoencoders

    Zou, Chenye and Jiao, Difan and Hu, Lijie. Deciphering Cultural Representations in Large Language Models via Sparse Autoencoders. Findings of the A ssociation for C omputational L inguistics: ACL 2026. 2026. doi:10.18653/v1/2026.findings-acl.278

  64. [72]

    H ist L ens: Mapping Idea Change across Concepts and Corpora

    Jing, Yi and Qiu, Weiyun and Peng, Yihang and Sui, Zhifang. H ist L ens: Mapping Idea Change across Concepts and Corpora. Proceedings of the 64th Annual Meeting of the A ssociation for C omputational L inguistics (Volume 1: Long Papers). 2026. doi:10.18653/v1/2026.acl-long.652

  65. [73]

    ParlaMint II: advancing comparable parliamentary corpora across Europe , volume =

    Erjavec, Tomaž and Kopp, Matyáš and Ljubešić, Nikola and Kuzman, Taja and Rayson, Paul and Osenova, Petya and Ogrodniczuk, Maciej and undefined\". ParlaMint II: advancing comparable parliamentary corpora across Europe , volume =. Language Resources and Evaluation , publisher =...

  66. [74]

    New Developments in the P olish Parliamentary Corpus

    Ogrodniczuk, Maciej and Nito \'n , Bart omiej. New Developments in the P olish Parliamentary Corpus. Proceedings of the Second ParlaCLARIN Workshop. 2020

  67. [75]

    The G erma P arl Corpus of Parliamentary Protocols

    Bl. The G erma P arl Corpus of Parliamentary Protocols. Proceedings of the Eleventh International Conference on Language Resources and Evaluation ( LREC 2018). 2018

  68. [76]

    2021 , eprint=

    Diachronic Analysis of German Parliamentary Proceedings: Ideological Shifts through the Lens of Political Biases , author=. 2021 , eprint=

  69. [77]

    Unfinished Business: Construction and Maintenance of a Semantically Tagged Historical Parliamentary Corpus, UK H ansard from 1803 to the present day

    Coole, Matthew and Rayson, Paul and Mariani, John. Unfinished Business: Construction and Maintenance of a Semantically Tagged Historical Parliamentary Corpus, UK H ansard from 1803 to the present day. Proceedings of the Second ParlaCLARIN Workshop. 2020

  70. [78]

    Italian Political Science Review/Rivista Italiana di Scienza Politica , author=

    A new database for Italian parliamentary speeches: introducing the ItaParlCorpus dataset , volume=. Italian Political Science Review/Rivista Italiana di Scienza Politica , author=. 2025 , pages=. doi:10.1017/ipo.2025.6 , number=

  71. [79]

    There ' s Something New about the I talian Parliament: The IPSA Corpus

    Frasnelli, Valentino and Palmero Aprosio, Alessio. There ' s Something New about the I talian Parliament: The IPSA Corpus. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

  72. [80]

    2018 , url=

    A Corpus of Grand National Assembly of Turkish Parliament’s Transcripts , author=. 2018 , url=

  73. [81]

    Manifesto Project Dataset , publisher =

    Volkens, Andrea and Lehmann, Pola and Matthieß, Theres and Merz, Nicolas and Regel, Sven and Weßels, Bernhard and. Manifesto Project Dataset , publisher =. doi:10.25522/MANIFESTO.MPDS.2017B , url =

  74. [82]

    2019 , copyright =

    Odell, Evan and Simon, Basile , title =. 2019 , copyright =. doi:10.5281/ZENODO.591264 , url =

  75. [83]

    2018 , eprint=

    Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change , author=. 2018 , eprint=

  76. [84]

    2019 , eprint=

    Survey of Computational Approaches to Lexical Semantic Change , author=. 2019 , eprint=

  77. [85]

    2020 , eprint=

    SemEval-2020 Task 1: Unsupervised Lexical Semantic Change Detection , author=. 2020 , eprint=

  78. [86]

    doi:10.13140/RG.2.2.14905.44649 , url =

    HistBERT: A Pre-trained Language Model for Diachronic Lexical Semantic Analysis , publisher =. doi:10.13140/RG.2.2.14905.44649 , url =

  79. [87]

    2019 , eprint=

    Cross-Lingual Alignment of Contextual Word Embeddings, with Applications to Zero-shot Dependency Parsing , author=. 2019 , eprint=

  80. [88]

    2025 , eprint=

    Framing Migration: A Computational Analysis of UK Parliamentary Discourse , author=. 2025 , eprint=

  81. [89]

    2023 , eprint=

    Sparse Autoencoders Find Highly Interpretable Features in Language Models , author=. 2023 , eprint=

  82. [90]

    2024 , eprint=

    Disentangling Dense Embeddings with Sparse Autoencoders , author=. 2024 , eprint=

  83. [91]

    2026 , eprint=

    DHPLT: large-scale multilingual diachronic corpora and word representations for semantic change modelling , author=. 2026 , eprint=

  84. [92]

    2025 , eprint=

    Narrative Media Framing in Political Discourse , author=. 2025 , eprint=

  85. [93]

    2026 , eprint=

    Cultural Adaptation in Large Language Models for Political Discourse , author=. 2026 , eprint=

  86. [94]

    2025 , eprint=

    Uncovering Cross-Linguistic Disparities in LLMs using Sparse Autoencoders , author=. 2025 , eprint=

  87. [95]

    2026 , eprint=

    Word-Centered Semantic Graphs for Interpretable Diachronic Sense Tracking , author=. 2026 , eprint=

  88. [96]

    2026 , eprint=

    SemEval-2026 Task 6: CLARITY -- Unmasking Political Question Evasions , author=. 2026 , eprint=

  89. [97]

    2026 , eprint=

    Cross-Architecture Model Diffing with Crosscoders: Unsupervised Discovery of Differences Between LLMs , author=. 2026 , eprint=

  90. [98]

    2026 , eprint=

    Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning , author=. 2026 , eprint=

  91. [99]

    2026 , eprint=

    Evolution of Concepts in Language Model Pre-Training , author=. 2026 , eprint=

  92. [100]

    2026 , eprint=

    Delta-Crosscoder: Robust Crosscoder Model Diffing in Narrow Fine-Tuning Regimes , author=. 2026 , eprint=

  93. [101]

    2026 , eprint=

    Transcoder Adapters for Reasoning-Model Diffing , author=. 2026 , eprint=

  94. [102]

    Crosscoding Through Time: Tracking Emergence & Consolidation Of Linguistic Representations Throughout LLM Pretraining

    Bayazit, Deniz and Mueller, Aaron and Bosselut, Antoine. Crosscoding Through Time: Tracking Emergence & Consolidation Of Linguistic Representations Throughout LLM Pretraining. Proceedings of the 64th Annual Meeting of the A ssociation for C omputational L inguistics (Volume 1:...

  95. [103]

    Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy

    Balagansky, Nikita and Aksenov, Yaroslav and Laptev, Daniil and Kurochkin, Vadim and Gerasimov, Gleb and Koriagin, Nikita and Gavrilov, Daniil. Train One Sparse Autoencoder Across Multiple Sparsity Budgets to Preserve Interpretability and Accuracy. Proceedings of the 2025 Conf...

  96. [104]

    Using relative entropy for detection and analysis of periods of diachronic linguistic change

    Degaetano-Ortlieb, Stefania and Teich, Elke. Using relative entropy for detection and analysis of periods of diachronic linguistic change. Proceedings of the Second Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Litera...

  97. [105]

    Grammar and Meaning: Analysing the Topology of Diachronic Word Embeddings

    Bizzoni, Yuri and Degaetano-Ortlieb, Stefania and Menzel, Katrin and Krielke, Pauline and Teich, Elke. Grammar and Meaning: Analysing the Topology of Diachronic Word Embeddings. Proceedings of the 1st International Workshop on Computational Approaches to Historical Language Ch...

  98. [106]

    Cross-lingual and Word-Independent Methods for Quantifying Degree of Grammaticalization

    Nagata, Ryo and Mochihashi, Daichi and Ido, Misato and Kubota, Yusuke and Otani, Naoki and Kawasaki, Yoshifumi and Takamura, Hiroya. Cross-lingual and Word-Independent Methods for Quantifying Degree of Grammaticalization. Proceedings of the 19th Conference of the E uropean Cha...

  99. [107]

    Transformer Circuits Thread , year =

    Jack Lindsey and Adly Templeton and Jonathan Marcus and Thomas Conerly and Joshua Batson and Christopher Olah , title =. Transformer Circuits Thread , year =

  100. [108]

    NeurIPS 2024 Workshop on Scientific Methods for Understanding Deep Learning , year =

    Bart Bussmann and Patrick Leask and Neel Nanda , title =. NeurIPS 2024 Workshop on Scientific Methods for Understanding Deep Learning , year =. 2412.06410 , archivePrefix =

  101. [109]

    Scaling and Evaluating Sparse Autoencoders , booktitle =

    Leo Gao and Tom Dupr. Scaling and Evaluating Sparse Autoencoders , booktitle =. 2025 , publisher =

  102. [110]

    Jumping Ahead: Improving Reconstruction Fidelity with

    Senthooran Rajamanoharan and Tom Lieberum and Nicolas Sonnerat and Arthur Conmy and Vikrant Varma and J. Jumping Ahead: Improving Reconstruction Fidelity with. 2024 , eprint =

  103. [111]

    Gemma Scope: Open Sparse Autoencoders Everywhere All at Once on

    Tom Lieberum and Senthooran Rajamanoharan and Arthur Conmy and Lewis Smith and Nicolas Sonnerat and Vikrant Varma and J. Gemma Scope: Open Sparse Autoencoders Everywhere All at Once on. 2024 , eprint =

  104. [112]

    2025 , eprint =

    Adam Karvonen and Can Rager and Johnny Lin and Curt Tigges and Joseph Bloom and David Chanin and Yeu-Tong Lau and Eoin Farrell and Callum McDougall and Kola Ayonrinde and Matthew Wearden and Arthur Conmy and Samuel Marks and Neel Nanda , title =. 2025 , eprint =

  105. [113]

    Pearce and Joseph Isaac Bloom and Curt Tigges and Noura Al Moubayed and Lee Sharkey and Neel Nanda , title =

    Patrick Leask and Bart Bussmann and Michael T. Pearce and Joseph Isaac Bloom and Curt Tigges and Noura Al Moubayed and Lee Sharkey and Neel Nanda , title =. The Thirteenth International Conference on Learning Representations (ICLR) , year =

  106. [114]

    2024 , eprint =

    Liv Gorton , title =. 2024 , eprint =

  107. [115]

    Proceedings of the 42nd International Conference on Machine Learning (ICML) , year =

    Subhash Kantamneni and Joshua Engels and Senthooran Rajamanoharan and Max Tegmark and Neel Nanda , title =. Proceedings of the 42nd International Conference on Machine Learning (ICML) , year =

  108. [116]

    Automatically Interpreting Millions of Features in Large Language Models , year =

    Gon. Automatically Interpreting Millions of Features in Large Language Models , year =. 2410.13928 , archivePrefix =

  109. [117]

    2024 , howpublished =

    Benjamin Wright and Lee Sharkey , title =. 2024 , howpublished =

  110. [118]

    2025 , eprint =

    Sabri Boughorbel and Fahim Dalvi and Nadir Durrani and Majd Hawasly , title =. 2025 , eprint =

  111. [119]

    Designing and Interpreting Probes with Control Tasks

    Hewitt, John and Liang, Percy. Designing and Interpreting Probes with Control Tasks. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019. doi:10.18...

  112. [120]

    2025 , eprint =

    Tatsuro Inaba and Kentaro Inui and Yusuke Miyao and Yohei Oseki and Benjamin Heinzerling and Yu Takagi , title =. 2025 , eprint =

  113. [121]

    Analysing Lexical Semantic Change with Contextualised Word Representations

    Giulianelli, Mario and Del Tredici, Marco and Fern \'a ndez, Raquel. Analysing Lexical Semantic Change with Contextualised Word Representations. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020. doi:10.18653/v1/2020.acl-main.365

  114. [122]

    and Leskovec, Jure and Jurafsky, Dan

    Hamilton, William L. and Leskovec, Jure and Jurafsky, Dan. Cultural Shift or Linguistic Drift? Comparing Two Computational Measures of Semantic Change. Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing. 2016. doi:10.18653/v1/D16-1229

  115. [123]

    Identifying Cross-Cultural Differences in Word Usage

    Garimella, Aparna and Mihalcea, Rada and Pennebaker, James. Identifying Cross-Cultural Differences in Word Usage. Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers. 2016

  116. [124]

    2023 , eprint =

    Stefano Montanelli and Francesco Periti , title =. 2023 , eprint =

  117. [125]

    ACM Computing Surveys , volume =

    Francesco Periti and Stefano Montanelli , title =. ACM Computing Surveys , volume =

  118. [126]

    How to Do Politics with Words: Investigating Speech Acts in Parliamentary Debates

    Reinig, Ines and Rehbein, Ines and Ponzetto, Simone Paolo. How to Do Politics with Words: Investigating Speech Acts in Parliamentary Debates. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

  119. [127]

    Target Based Speech Act Classification in Political Campaign Text

    Subramanian, Shivashankar and Cohn, Trevor and Baldwin, Timothy. Target Based Speech Act Classification in Political Campaign Text. Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM 2019). 2019. doi:10.18653/v1/S19-1030

  120. [128]

    PUB : A Pragmatics Understanding Benchmark for Assessing LLM s' Pragmatics Capabilities

    Sravanthi, Settaluri Lakshmi and Doshi, Meet and Tankala, Pavan and Murthy, Rudra and Dabre, Raj and Bhattacharyya, Pushpak. PUB : A Pragmatics Understanding Benchmark for Assessing LLM s' Pragmatics Capabilities. Findings of the Association for Computational Linguistics: ACL ...

  121. [129]

    Using U niversal D ependencies in Cross-Linguistic Complexity Research

    Berdicevskis, Aleksandrs and. Using U niversal D ependencies in Cross-Linguistic Complexity Research. Proceedings of the Second Workshop on Universal Dependencies (UDW 2018). 2018. doi:10.18653/v1/W18-6002

  122. [130]

    A Framework for Understanding the Role of Morphology in U niversal D ependency Parsing

    Dehouck, Mathieu and Denis, Pascal. A Framework for Understanding the Role of Morphology in U niversal D ependency Parsing. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018

  123. [131]

    U niversal D ependencies: A Cross-Linguistic Perspective on Grammar and Lexicon

    Nivre, Joakim. U niversal D ependencies: A Cross-Linguistic Perspective on Grammar and Lexicon. Proceedings of the Workshop on Grammar and Lexicon: Interactions and Interfaces (GramLex). 2016

  124. [132]

    Harald Baayen , title =

    R. Harald Baayen , title =. Corpus Linguistics: An International Handbook , editor =

  125. [133]

    Laurie Bauer , title =

  126. [134]

    Bergen Language and Linguistics Studies , volume =

    Marie-Pauline Krielke , title =. Bergen Language and Linguistics Studies , volume =

  127. [135]

    2022 , eprint =

    Konstantin Todorov and Giovanni Colavizza , title =. 2022 , eprint =

  128. [136]

    Proceedings of the 12th International Conference on Agents and Artificial Intelligence (ICAART) , year =

    Daniel van Strien and Kaspar Beelen and Mariona Coll Ardanuy and Kasra Hosseini and Barbara McGillivray and Giovanni Colavizza , title =. Proceedings of the 12th International Conference on Agents and Artificial Intelligence (ICAART) , year =

  129. [137]

    Hill and Simon Hengchen , title =

    Mark J. Hill and Simon Hengchen , title =. Digital Scholarship in the Humanities , volume =

  130. [138]

    Joint Semantic Synthesis and Morphological Analysis of the Derived Word

    Cotterell, Ryan and Sch. Joint Semantic Synthesis and Morphological Analysis of the Derived Word. Transactions of the Association for Computational Linguistics. 2018. doi:10.1162/tacl_a_00003

  131. [139]

    MistralAI , journal=

  132. [140]

    Why do language models perform worse for morphologically complex languages?

    Arnett, Catherine and Bergen, Benjamin. Why do language models perform worse for morphologically complex languages?. Proceedings of the 31st International Conference on Computational Linguistics. 2025

  133. [141]

    A Morphology-Aware Evaluation of T urkish Syntax in Large Language Models

    Ba s ar, Ezgi and Bisazza, Arianna. A Morphology-Aware Evaluation of T urkish Syntax in Large Language Models. Proceedings of the Second Workshop Natural Language Processing for T urkic Languages ( SIGTURK 2026). 2026. doi:10.18653/v1/2026.sigturk-1.9

  134. [143]

    Proceedings of the National Academy of Sciences , volume =

    Futrell, Richard and Mahowald, Kyle and Gibson, Edward , title =. Proceedings of the National Academy of Sciences , volume =. 2015 , doi =

  135. [144]

    and Dautriche, Isabelle and Mahowald, Kyle and Bergen, Leon and Levy, Roger , title =

    Gibson, Edward and Futrell, Richard and Piantadosi, Steven T. and Dautriche, Isabelle and Mahowald, Kyle and Bergen, Leon and Levy, Roger , title =. Trends in Cognitive Sciences , volume =. 2019 , doi =

  136. [145]

    and Gibson, Edward , title =

    Futrell, Richard and Levy, Roger P. and Gibson, Edward , title =. Language , volume =. 2020 , doi =

  137. [146]

    Proceedings of the National Academy of Sciences , volume =

    Hahn, Michael and Xu, Yang , title =. Proceedings of the National Academy of Sciences , volume =. 2022 , doi =

  138. [147]

    Our kind of people? D etecting populist references in political debates

    Klamm, Christopher and Rehbein, Ines and Ponzetto, Simone Paolo. Our kind of people? D etecting populist references in political debates. Findings of the Association for Computational Linguistics: EACL 2023. 2023. doi:10.18653/v1/2023.findings-eacl.91

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.