Pith. sign in

REVIEW 4 major objections 4 minor 32 references

Let the Fuzzy Rule Speak: Enhancing In-context Learning Debiasing with Interpretability

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Fuzzy rules cut LLM class-accuracy bias by 56%

desk verdict FuRud is a plausible incremental extension of DNIP, but the headline 21% gain is driven almost entirely by one outlier dataset (DDI); without it, FuRud no longer beats DNIP on average accuracy. read the letter →

arxiv 2412.19018 v4 pith:EKH73JTP submitted 2024-12-26 cs.CL

Ruixi Lin , Yang You This is my paper · ORCID
classification cs.CL
keywords fuzzyrulein-contextlearningdebiasingclassimbalanceinterpretabilityprobabilitycorrectiontriangularmembershipfunctionsCOBias
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

In-context learning (ICL) lets large language models classify text by prompting, but per-class accuracy is often lopsided: some classes are over-predicted while others are under-predicted. The paper claims this imbalance can be fixed at inference time, without retraining or updating the model, by transforming each class's output probability through a range-specific correction curve. The curves are triangular membership functions from fuzzy logic, and each class picks one of 19 candidates by solving a small combinatorial optimization on labeled examples. On seven classification benchmarks, the method, called FuRud, cuts the average pairwise class-accuracy gap (COBias) by 56% and raises overall accuracy from 59.4% to 72.0%. If right, this gives an interpretable, cheap way to debias ICL outputs.

What carries the argument

The central machinery is a family of 19 triangular membership functions, each a piecewise-linear map f(p) defined by three parameters (a, b, c), together with a per-class integer selection variable κ_i that picks one map per output class. Each map transforms a class probability according to which probability range it falls into, allowing asymmetric amplification and reduction. A multi-objective integer program combines three objectives—minimize COBias, maximize overall accuracy, and penalize classes below an accuracy threshold—into a single energy function that is minimized by simulated annealing on a labeled optimization set. At inference, the learned maps are applied to each test instance's class probabilities and the final class is the argmax of the corrected probabilities.

What would settle it

If, on a held-out split of a benchmark, the per-class accuracy within probability ranges is inverted or flat compared with the optimization set—for instance, low-probability instances turn out to be more accurate than high-probability ones—then the learned membership functions would amplify the wrong ranges and FuRud's reported gains should disappear. A direct check is to recompute FuRud's corrections on a dataset where the optimization set and test set come from different distributions or time periods and see whether accuracy and COBias revert toward or below the un-corrected ICL baseline.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the accuracy imbalance in ICL classifiers is range-dependent: within a ground-truth class, instances whose model-assigned probability falls in a low range tend to be misclassified, while instances with higher probability in that class are usually correct. Consequently, a single class-level weight—as used by prior debiasing—misses this structure, and the right correction is a piecewise-linear map that amplifies the low range, sometimes leaves middle ranges alone, and can shrink high ranges of competing classes. FuRud learns one such map per class from a labeled optimization set, choosing among 19 triangular membership functions by minimizing a weighted sum of class-accuracy bias (COBias), overall error, and an extreme-class penalty. The selected maps then transform test-time probabilities, and a 'Don't Change' identity function is selected when a class is already accurate. The paper reports average test accuracy rising from 59.4% to 72.0% and average COBias falling from 40.5% to 17.8% across AGNews, DBpedia, SST-5, TREC, RTE, DDI, and PubMedQA.

Load-bearing premise

The load-bearing premise is that the relationship between the magnitude of an in-context output probability for a class and the true accuracy of instances in that probability range is stable enough to transfer from the labeled optimization set to the unlabeled test distribution.

Editorial extensions

If this is right

  • FuRud gives a post-hoc, training-free debiasing route: no LLM parameter updates, only a few minutes of CPU search, so it applies to models whose weights are frozen or inaccessible.
  • The selected membership functions double as an explanation: a class assigned 'Don't Change' is already accurate, while an amplified low-range or shrunk high-range points to where the model's probability scale is miscalibrated.
  • The method transfers to harder prompting setups: under letter-answer prompts it raises average accuracy by 44% relative and cuts COBias by 54% relative, and under k-shot demonstrations it raises average accuracy to 73.5%.
  • FuRud works with as few as 10 labeled optimization examples, suggesting it can be deployed quickly on new tasks without collecting large validation sets.
  • The accuracy-COBias tradeoff is controllable through the granularity of the fuzzy partitions, with mixed partitions giving the best joint result.

Reading between the lines

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

  • Because the method operates purely on output probabilities, the same machinery could be applied to any probabilistic classifier—vision models, rerankers, or regression-to-class reductions—whenever the range-accuracy monotonicity holds, not just to decoder-only LLMs.
  • The interpretability output could be used as a diagnostic for prompt design: if a class frequently selects an amplification function, the prompt's demonstrations or verbalizer are likely biasing that class down, guiding targeted prompt edits.
  • A natural stress test would be to learn membership functions on one domain and apply them to another, to see whether the selected fuzzy rules are generalizable calibration priors or overfit to the optimization set.
  • The 'Don't Change' selection rate could itself serve as a metric of how imbalanced a task is, giving a probe into model bias before debiasing.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes FuRud, a post-hoc debiasing method for in-context learning (ICL) text classification. For each class, FuRud selects one of 19 triangular fuzzy membership functions by solving a weighted multi-objective integer program via simulated annealing on a labeled optimization set, then applies the selected functions to test-time class probabilities to reduce class-accuracy bias (COBias) and improve overall accuracy. Experiments on seven benchmarks with Llama-2-13B report average accuracy rising from 59.4% to 72.0% and COBias falling from 40.5% to 17.8%, with additional results on two more models, a few-shot optimization setting, and letter-based prompts.

Significance. The idea of using fuzzy membership functions to provide interpretable, range-specific probability corrections is novel and practically attractive because it requires no LLM fine-tuning and operates purely at inference time. The paper follows reasonable experimental hygiene: separate optimization/development/test splits, three runs per dataset, and supplementary analyses on additional models and optimization sizes. If the reported improvements were robust, FuRud would be a useful addition to the ICL debiasing toolbox. However, the headline claim of outperforming state-of-the-art methods is not robust to the removal of the DDI dataset, and the COBias comparison to DNIP is not favorable. The central empirical claim therefore needs re-examination and reframing.

major comments (4)
  1. [Table 1, Abstract] The headline claim that FuRud 'outperforms state-of-the-art debiasing methods across seven benchmarks' is not robust to the DDI outlier. In Table 1, DDI has an ICL accuracy of 7.2% (far below random for a 5-class task) and FuRud raises it to 69.3%, a 62.1-point jump. Excluding DDI from the seven-dataset averages, ICL accuracy becomes 68.1% and FuRud becomes 72.4%, a relative improvement of only 6.4% rather than the claimed 21%; DNIP's average without DDI is 74.0%, so FuRud no longer outperforms DNIP. Per-dataset, DNIP is higher on AGNews, DBpedia, and PubMedQA. The paper gives no analysis of why DDI's ICL baseline is so low or why FuRud succeeds there, so the cross-benchmark claim is not established.
  2. [Table 1, Section 4.2] The COBias comparison to DNIP is unfavorable and should be reported honestly. FuRud's average COBias is 17.8% versus DNIP's 14.3%, and DNIP achieves lower COBias on six of the seven datasets (AGNews, DBpedia, SST-5, TREC, RTE, DDI). The abstract's 'outperforming state-of-the-art' is therefore only defensible on accuracy, and even that claim depends on the DDI outlier discussed above.
  3. [Section 1, Figure 4] The paper's motivating assumption that, within a ground-truth class, instances with lower ICL probability have lower accuracy is asserted in Section 1 as a 'general observation' but supported only by a single example in Figure 4 (class Business on AGNews). Since the interpretability and range-specific correction claims rest on this relationship, the authors should provide a systematic validation across all classes and datasets, for example by reporting the correlation between probability bins and per-bin accuracy before correction.
  4. [Section 4.1] The hyperparameters of the optimization are not reported. The paper says α and β are 'tuned on the development set' and λ is set to 0.5, but the actual values of α and β, the simulated annealing schedule, the number of iterations, and the number of restarts are missing. Because the method is an integer-programming heuristic with a search space of 19^N, these details are needed to assess stability and reproducibility.
minor comments (4)
  1. [Table 1] The table header contains the typo 'PubMedaQA'; it should be 'PubMedQA'.
  2. [Throughout] The acronym is written inconsistently as 'FuRuD' (Section 4.2) and 'FuRud' elsewhere; please unify.
  3. [Equation (3)] Equation (3) is self-referential: p'_mi is defined using the sum of p'_mi on the right-hand side. Please clarify the intended computation order, e.g., first compute each f_k(p_mi) for the selected κ_i, then reset to the original probability if all are zero.
  4. [Section 5.2] There is a typo in the sentence 'although he DC partition alone can obtain 15% higher accuracy'; 'he' should be 'the'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: FuRud's test-set gains are genuine predictions from a separate optimization set.

full rationale

FuRud's derivation chain is a standard supervised learning loop: ICL class probabilities on a labeled optimization set are used to select membership functions under the explicit objectives in Eqs. 4-6, and the selected functions are then applied to a disjoint test set. The reported accuracy and COBias numbers are therefore predictive, not fitted values; optimizing the same metrics on the optimization set does not force the test-set result. The COBias metric is defined in the paper itself (Eq. 4), so it does not depend on an invisible imported definition. The 19 triangular membership functions are a stated design choice (Section 3, Figure 2), not derived from the target result. The reuse of the authors' prior DNIP work for simulated annealing implementation and dataset preprocessing is methodological and non-load-bearing: no uniqueness theorem or external authority is invoked to rule out alternative correction functions, and DNIP is treated as a baseline rather than as the source of the paper's correctness. The sensitivity of the headline improvement to the DDI dataset is a robustness concern about generalization, not a circularity in the derivation. No step reduces by construction to its own inputs, so no circular step is identified.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The method is entirely empirical: it fits correction functions on an optimization set. The main free parameters are the objective weights and threshold. No new theoretical entities are introduced, and the causal claims about probability ranges rest on unverified domain assumptions.

free parameters (4)
  • alpha (COBias weight) = not reported
    Hyperparameter in the energy function Z = 1 - Z_Acc + alpha * Z_COBias + beta * Z_Extreme; tuned on the development set, values not given in the paper.
  • beta (extreme accuracy weight) = not reported
    Same as alpha, tuned on the development set.
  • lambda (accuracy threshold) = 0.5
    Set to a fixed value for simplicity; not tuned.
  • membership function parameters (a_k, b_k, c_k) = fixed library of 19
    Hand-designed grid of triangular functions; defines the search space, not fitted to data.
assumptions (4)
  • domain assumption For each class, instances with lower ICL output probability for the true class have lower accuracy than instances with higher probability.
    Stated in Section 1 as a 'general observation' but not empirically quantified; the entire range-specific correction relies on this relationship being stable between optimization and test sets.
  • domain assumption The optimization set is representative of the test distribution for the learned membership functions.
    Standard supervised assumption; the paper splits data but does not characterize distribution shift.
  • domain assumption Simulated annealing finds a sufficiently good solution to the nonlinear integer program.
    SA is a heuristic; no optimality guarantee, details are delegated to prior work.
  • domain assumption The family of 19 triangular functions is expressive enough to capture beneficial corrections.
    No analysis of the function space; the ablation in Section 5.2 partially explores granularity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Let the Fuzzy Rule Speak: Enhancing In-context Learning Debiasing with Interpretability." pith.science (2026). https://pith.science/paper/EKH73JTP

@misc{pith2026241219018,
  author       = {Pith},
  title        = {Pith review of: Let the Fuzzy Rule Speak: Enhancing In-context Learning Debiasing with Interpretability},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EKH73JTP}},
  note         = {Machine review of arXiv:2412.19018}
}
read the original abstract

Large language models (LLMs) often struggle with balanced class accuracy in text classification tasks using in-context learning (ICL), hindering some practical uses due to user dissatisfaction or safety risks caused by misclassifications. Retraining LLMs to address root causes in data or model priors is neither easy nor cost-effective. This paper delves deeper into the class accuracy imbalance issue, identifying that it arises because certain classes consistently receive disproportionately high ICL probabilities, causing under-prediction and lower accuracy for others. More importantly, probability ranges affect the imbalance differently, allowing for precise, range-specific corrections. We introduce FuRud (Fuzzy Rule Optimization-based Debiasing), a method for sample-level class probability correction. FuRud tackles interpretability challenges by determining why certain classes need corrections and tailoring adjustments for each instance's class probabilities which is powered by fuzzy sets with triangular membership functions, transforming a class probability based on the range it belongs to. By solving a nonlinear integer programming problem with a labeled set of ICL class probabilities to minimize class accuracy bias (COBias) and maximize overall accuracy, each class selects an optimal correction function from 19 triangular membership functions without updating an LLM, and the selected functions correct test instances at inference. Across seven benchmark datasets, FuRud reduces COBias by over half (56%) and improves overall accuracy by 21% relatively, outperforming state-of-the-art debiasing methods.

Figures

Figures reproduced from arXiv: 2412.19018 by the authors.

Figure 1
Figure 1. An overview of FuRud. ICL output probabilities across answer classes for input instances [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. 19 triangular membership functions. Membership Functions. Each class selects from 19 triangular membership functions. Tri￾angular membership functions are popular for fuzzy rule-based classification [8], as the speed of changes is easily controlled by the slope, and the linearity is computationally efficient. With￾out knowing the expected fuzzy partitions in downstream datasets, we employ four fuzzy par￾titions, res… view at source ↗
Figure 3
Figure 3. Class probabilities before and after applying corrections. For each task, we report results [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Quantitative evaluation on the ratio of instances that benefit from the correction, exemplified [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Accuracy-COBias tradeoff of fuzzy partitions. As shown by [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Few-shot optimization. We additionally prompt Llama-2-13B with the following demonstration selection strategy: k-shot prompting, where k is the number of classes. A demonstrative example from each class is randomly selected from the optimiza￾tion set and represented in…
Figure 7
Figure 7. Figure 7: Additional few-shot optimization results. [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 27 canonical work pages

  1. [1]

    Cobias and debias: Minimizing language model pairwise accuracy bias via nonlinear integer programming, 2024

    Ruixi Lin and Yang You. Cobias and debias: Minimizing language model pairwise accuracy bias via nonlinear integer programming, 2024. URL https://arxiv.org/abs/2405.07623

    arXiv 2024

  2. [2]

    Have We Learned to Explain?: How Interpretability Methods Can Learn to Encode Predictions in their Interpretations

    Neil Jethani, Mukund Sudarshan, Yindalon Aphinyanaphongs, and Rajesh Ranganath. Have We Learned to Explain?: How Interpretability Methods Can Learn to Encode Predictions in their Interpretations. Proceedings of Machine Learning Research, 130:1459–1467, 2021. URL https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8096519

    work page 2021

  3. [3]

    Carvalho, Eduardo M

    Diogo V . Carvalho, Eduardo M. Pereira, and Jaime S. Cardoso. Machine learning interpretability: A survey on methods and metrics. Electronics, 8(8), 2019. URL https://www.mdpi.com/ 2079-9292/8/8/832

    work page 2019

  4. [4]

    Vernon, Naoki Masuyama, and Yusuke Nojima

    Eric M. Vernon, Naoki Masuyama, and Yusuke Nojima. Integrating white and black box techniques for interpretable machine learning. In Xin-She Yang, Simon Sherratt, Nilanjan Dey, and Amit Joshi, editors, Proceedings of Ninth International Congress on Information and Communication Technology, pages 639–649, 2024

    work page 2024

  5. [5]

    Explainable artificial intelligence: a systematic review, 2020

    Giulia Vilone and Luca Longo. Explainable artificial intelligence: a systematic review, 2020. URL https://arxiv.org/abs/2006.00093

    arXiv 2020

  6. [6]

    Analysis of interpretability-accuracy tradeoff of fuzzy systems by multiobjective fuzzy genetics-based machine learning

    Hisao Ishibuchi and Yusuke Nojima. Analysis of interpretability-accuracy tradeoff of fuzzy systems by multiobjective fuzzy genetics-based machine learning. International Journal of Approximate Reasoning, 44(1):4–31, 2007. URL https://www.sciencedirect.com/ science/article/pii/S0888613X06000405

    work page 2007

  7. [7]

    Performance Evaluation of Fuzzy Classifier Systems for Multidimensional Pattern Classification Problems

    Hisao Ishibuchi, Tomoharu Nakashima, and Tadahiko Murata. Performance Evaluation of Fuzzy Classifier Systems for Multidimensional Pattern Classification Problems. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 29(5):601–618, 1999. doi: 10.1109/3477. 790443. URL https://ieeexplore.ieee.org/document/790443. 9

    doi:10.1109/3477 1999

  8. [8]

    Hybridization of Fuzzy GBML Approaches for Pattern Classification Problems

    Hisao Ishibuchi, Takashi Yamamoto, and Tomoharu Nakashima. Hybridization of Fuzzy GBML Approaches for Pattern Classification Problems. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 35(2):359–365, 2005. URL https://ieeexplore.ieee. org/abstract/document/1408064

    work page arXiv 2005

Show all 32 references
  1. [9]

    Multiobjective Fuzzy Genetics-based Machine Learning with a Reject Option

    Yusuke Nojima and Hisao Ishibuchi. Multiobjective Fuzzy Genetics-based Machine Learning with a Reject Option. In 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), pages 1405–1412, 2016

    2016

  2. [10]

    A multi-objective genetic optimization of interpretability-oriented fuzzy rule-based classifiers

    Filip Rudzi ´nski. A multi-objective genetic optimization of interpretability-oriented fuzzy rule-based classifiers. Applied Soft Computing , 38:118–133, 2016. URL https://www. sciencedirect.com/science/article/abs/pii/S1568494615006109

    2016

  3. [11]

    Gorzałczany and Filip Rudzi ´nski

    Marian B. Gorzałczany and Filip Rudzi ´nski. Interpretable and accurate medical data classi- fication – a multi-objective genetic-fuzzy optimization approach. Expert Systems with Ap- plications, 71:26–39, 2017. doi: https://doi.org/10.1016/j.eswa.2016.11.017. URL https: //www....

    2017 doi

  4. [12]

    L.A. Zadeh. Fuzzy sets. Information and Control, 8(3):338–353, 1965. URL https://www. sciencedirect.com/science/article/pii/S001999586590241X

    1965

  5. [13]

    Language models are few-shot learners

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-V oss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel Ziegler, Jeffrey ...

    1901

  6. [14]

    Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP

    Timo Schick, Sahana Udupa, and Hinrich Schütze. Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP. Transactions of the Association for Com- putational Linguistics, 9:1408–1424, 12 2021. URL https://doi.org/10.1162/tacl_a_ 00434

    2021 doi

  7. [15]

    Calibrate Before Use: Improving Few-shot Performance of Language Models

    Zihao Zhao, Eric Wallace, Shi Feng, Dan Klein, and Sameer Singh. Calibrate Before Use: Improving Few-shot Performance of Language Models. InProceedings of the 38th International Conference on Machine Learning, pages 12697–12706, 2021. URL https://proceedings. mlr.press/v139/zh...

    2021

  8. [16]

    Sewon Min, Xinxi Lyu, Ari Holtzman, Mikel Artetxe, Mike Lewis, Hannaneh Hajishirzi, and Luke Zettlemoyer. Rethinking the Role of Demonstrations: What Makes In-Context Learning Work? In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, page...

    2022

  9. [17]

    Surface Form Competition: Why the Highest Probability Answer Isn’t Always Right

    Ari Holtzman, Peter West, Vered Shwartz, Yejin Choi, and Luke Zettlemoyer. Surface Form Competition: Why the Highest Probability Answer Isn’t Always Right. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 7038–7051,

    2021

  10. [18]

    Exploiting cloze-questions for few-shot text classifica- tion and natural language inference

    Timo Schick and Hinrich Schütze. Exploiting cloze-questions for few-shot text classifica- tion and natural language inference. In Paola Merlo, Jorg Tiedemann, and Reut Tsarfaty, editors, Proceedings of the 16th Conference of the European Chapter of the Association for Computat...

    2021

  11. [19]

    Mitigating label biases for in- context learning

    Yu Fei, Yifan Hou, Zeming Chen, and Antoine Bosselut. Mitigating label biases for in- context learning. In Proceedings of the 61st Annual Meeting of the Association for Com- putational Linguistics (Volume 1: Long Papers) , pages 14014–14031, July 2023. URL https://aclanthology...

    2023

  12. [20]

    Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering

    Han Zhou, Xingchen Wan, Lev Proleev, Diana Mincu, Jilin Chen, Katherine A Heller, and Subhrajit Roy. Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering. In The Twelfth International Conference on Learning Representations, 2024. URL https:/...

    2024

  13. [21]

    Garey and D.S

    M.R. Garey and D.S. Johnson. Computers and Intractability: A Guide to the Theory of NP- completeness. Mathematical Sciences Series. Freeman, 1979. ISBN 9780716710448. URL https://books.google.com.sg/books?id=fjxGAQAAIAAJ

    1979

  14. [22]

    Character-level Convolutional Net- works for Text Classification

    Xiang Zhang, Junbo Zhao, and Yann LeCun. Character-level Convolutional Net- works for Text Classification. In Advances in Neural Information Processing Sys- tems, 2015. URL https://proceedings.neurips.cc/paper_files/paper/2015/ file/250cf8b51c773f3f8dc8b4be867a9a02-Paper.pdf

    2015

  15. [23]

    DBpedia: A Nucleus for A Web of Open Data

    Sören Auer, Christian Bizer, Georgi Kobilarov, Jens Lehmann, Richard Cyganiak, and Zachary Ives. DBpedia: A Nucleus for A Web of Open Data. In Proceedings of the 6th International The Semantic Web and 2nd Asian Conference on Asian Semantic Web Conference, pages 722–735, 2007

    2007

  16. [24]

    Manning, Andrew Ng, and Christopher Potts

    Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Christopher D. Manning, Andrew Ng, and Christopher Potts. Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Proces...

    2013

  17. [25]

    V oorhees and Dawn M

    Ellen M. V oorhees and Dawn M. Tice. Building a question answering test collection. InProceed- ings of the 23rd Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 200–207, 2000. URL https://doi.org/10.1145/345508. 345577

    doi

  18. [26]

    Learning Question Classifiers

    Xin Li and Dan Roth. Learning Question Classifiers. In COLING 2002: The 19th Interna- tional Conference on Computational Linguistics, 2002. URL https://aclanthology.org/ C02-1150

    2002

  19. [27]

    The PASCAL Recognising Textual Entailment Challenge

    Ido Dagan, Oren Glickman, and Bernardo Magnini. The PASCAL Recognising Textual Entailment Challenge. In Machine Learning Challenges. Evaluating Predictive Uncertainty, Visual Object Classification, and Recognising Tectual Entailment, pages 177–190, 2006

    2006

  20. [28]

    SemEval-2013 Task 9 : Extraction of Drug-Drug Interactions from Biomedical Texts (DDIExtraction 2013)

    Isabel Segura-Bedmar, Paloma Martínez, and María Herrero-Zazo. SemEval-2013 Task 9 : Extraction of Drug-Drug Interactions from Biomedical Texts (DDIExtraction 2013). In Second Joint Conference on Lexical and Computational Semantics (*SEM), Volume 2: Proceedings of the Seventh ...

    2013

  21. [29]

    PubMedQA: A Dataset for Biomedical Research Question Answering

    Qiao Jin, Bhuwan Dhingra, Zhengping Liu, William Cohen, and Xinghua Lu. PubMedQA: A Dataset for Biomedical Research Question Answering. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Con- ference on Nat...

    2019

  22. [30]

    Chain-of-thought unfaithfulness as disguised accuracy, 2024

    Oliver Bentham, Nathan Stringham, and Ana Marasovi´c. Chain-of-thought unfaithfulness as disguised accuracy, 2024. URL https://arxiv.org/abs/2402.14897. 11 A Details on Membership Functions Table 4 lists the details about the membership functions used in this work. Function Pa...

    2024 arXiv

  23. [2013]

    URL https://aclanthology.org/S13-2056.pdf

  24. [2021]

    URL https://aclanthology.org/2021.emnlp-main.564

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.