Pith. sign in

REVIEW 4 major objections 6 minor 37 references

Data Quality Enhancement on the Basis of Diversity with Large Language Models for Text Classification: Uncovered, Difficult, and Noisy

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read A half-size training set, selected for diversity, beats full-data fine-tuning on every benchmark tested.

desk verdict A practical half-data selection recipe with a plausible but not yet solid empirical claim; the idea is good, the evidence needs seeds and code. read the letter →

arxiv 2412.06575 v2 pith:I76FJTUE submitted 2024-12-09 cs.CL

classification cs.CL
keywords dataqualityenhancementtextclassificationlargelanguagemodelsfine-tuningselectionlabelnoisediversitysamplinguncovereddifficultnoisy
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

The paper proposes a data quality enhancement (DQE) method for fine-tuning large language models on text classification. It first uses K-Center-Greedy sampling to pick a diverse half of the training set, fine-tunes on that half, and then predicts the unsampled half. Incorrectly predicted samples are split into three categories using cosine similarity to their nearest neighbor: uncovered, difficult, and noisy. Noisy samples are removed while uncovered and difficult samples are added to the final set, which stays around half the original size. The paper claims this DQE-selected set beats full-data fine-tuning on all six public benchmarks while saving nearly half the training time.

What carries the argument

The load-bearing mechanism is the three-way classification of incorrectly predicted samples based on the maximum cosine similarity to any other training sample. A sample is 'uncovered' if its nearest neighbor lies in the unsampled set and shares its label, 'difficult' if the nearest neighbor lies in the sampled set and shares its label, and 'noisy' if the nearest neighbor carries a different label. This single-nearest-neighbor rule, applied on top of K-Center-Greedy diversity sampling, determines which mispredicted samples are added back to the training set and which are removed. The 'noisy' candidates are further confirmed by GPT-4o before deletion.

What would settle it

Evaluate DQE on a dataset whose labels are hand-verified to be noise-free, or on a dataset with artificially injected label noise at known rates; if the method still removes a large proportion of clean examples or fails to recover the injected noise, then the nearest-neighbor disagreement heuristic is not isolating label noise, and the accuracy gains should be attributed to something other than noise removal.

Watch

Extended reading notes

Core claim

The central discovery is that a training set built from a diversity-greedy half, enriched with mispredicted samples whose nearest neighbor shares their label, and purged of samples whose nearest neighbor disagrees, yields higher test accuracy than training on the whole dataset. The authors attribute this to label noise: full-data fine-tuning is dragged down by mislabeled examples, which the greedy seed set partially avoids and DQE further removes. Across MR, CR, IMDb, SST-2, SST-5, and AG News, the DQE-selected model achieves the highest accuracy in every task, outperforming full-data training by margins up to 1.13 points, with statistically significant gains in four of six datasets. The paper also shows that some remaining test-set errors are themselves due to incorrect labels, so the true performance gap may be even wider.

Load-bearing premise

The method assumes that any two highly similar text items should have the same label, so a nearest neighbor with a different label is treated as a labeling error; if that assumption fails for near-boundary or multi-faceted examples, valid data get discarded and the method's gains could reverse.

Editorial extensions

If this is right

  • Fine-tuning an LLM for text classification can use roughly half the training data and still outperform full-data training, cutting training time approximately in half.
  • The three-way error taxonomy gives a practical recipe for cleaning noisy public benchmarks without full manual re-annotation.
  • The method is model-agnostic and scalable: it needs only a vectorizer, a greedy sampler, and one fine-tuned model, so it applies to large datasets.
  • Because some test-set errors are themselves mislabeled, DQE models' true accuracy may be higher than the reported numbers on noisy test sets.

Reading between the lines

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

  • The nearest-neighbor rule may be too aggressive on genuinely hard or near-boundary examples where similar texts legitimately carry different labels, so the method could discard valid data; a multi-neighbor vote or similarity threshold might improve robustness.
  • The noise-verification step depends on GPT-4o, which is a reproducibility concern; the paper does not test whether a smaller open-source verifier would give the same results.
  • Combining DQE with active learning or iterative re-sampling could use the 'uncovered' signal to guide further data collection, potentially extending the gains beyond a single pass.
  • The data-efficiency gain suggests that within a fixed compute budget one could double the model size or training epochs while keeping the same cost, potentially yielding even higher accuracy than reported.
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 manuscript proposes DQE, a data-selection pipeline for fine-tuning large language models on text classification. The method first selects a diverse half of the training set using K-Center-Greedy on embeddings from all-mpnet-base-v2, fine-tunes Qwen2-7B-Instruct on that subset, and then labels misclassified unsampled examples as uncovered, difficult, or noisy by comparing each to its nearest neighbor in the full training set via cosine similarity. Uncovered and difficult examples are added to the final training set, while noisy examples are verified by GPT-4o and removed. On six public benchmarks the authors report that the resulting roughly half-size training set yields higher test accuracy than full-data fine-tuning and also surpasses greedy sampling on every task, leading them to claim better accuracy with nearly half the training time.

Significance. If established, the claim that a carefully selected half-size training set can outperform the full dataset for LLM fine-tuning would be practically important, and the paper's decomposition of selected data into uncovered, difficult, and noisy categories is a useful interpretive framing. The authors use standard public datasets and an open base model, and the core comparison across Full-Data, Greedy, and DQE is performed with the same base model and hyperparameters, which is a reasonable experimental design. However, the evidence is currently insufficient to support the central claim: all results come from single fine-tuning runs with no seed variation, the significance test in Table 3 is applied to per-sample predictions of two fixed models rather than across independent runs, and the load-bearing noise-detection assumption in Section 3.3 is not validated quantitatively. The absolute gains over Full-Data are small (0.32 to 1.13 percentage points), so without variance estimates they are consistent with training noise.

major comments (4)
  1. [§4.4, Table 3] The statistical test does not support the claim that DQE reliably outperforms Full-Data. The t-test is computed on per-sample binary outcomes (1 for correct, 0 for incorrect) for two fixed models, so it can only describe a difference on this particular test set for these particular trained models. It does not account for variance across independent fine-tuning runs, and the paper reports only one run per condition. To establish the central claim, the authors should run multiple random seeds for each method and dataset, report mean accuracy and standard deviation, and use an appropriate test on the distribution of accuracy differences across runs.
  2. [§3.3, Figure 4] The noise-detection step relies on the premise that two highly similar texts must have the same label, but this premise is not justified and is likely violated for legitimate near-boundary examples, especially in sentiment classification and AG News. The generic all-mpnet-base-v2 embedding is not task-specific, so valid opposite-label texts can have high cosine similarity. The paper provides only a handful of anecdotal examples in Figure 4 and an unmeasured GPT-4o verification, with no precision or recall analysis of the noise detector against human-annotated labels. Without this validation, the claim that DQE enhances quality by removing label noise is not established; the removed examples could be valid training data, and the reported accuracy gains could reflect a different selection effect.
  3. [§4.4, Table 2] The reported improvements of DQE over Full-Data are small (0.32% for AG News, 0.33% for SST-5, 0.53% for CR, 0.54% for IMDb, 0.88% for SST-2, 1.13% for MR) and are presented without any measure of run-to-run variability. With a single training run per condition, these differences are within the range that seed variation can produce. Additionally, the paper does not provide a significance test for DQE versus Greedy, even though the narrative that DQE improves on greedy sampling is supported only by the point estimates in Table 2.
  4. [§4.4, 'Baseline' column in Table 2] The comparison against the PapersWithCode baselines is not a controlled comparison: those numbers come from different model architectures, training procedures, and hyperparameter choices, so the statement that DQE 'has achieved state-of-the-art performance' is not supported by direct evidence. The same-setting comparison among Full-Data, Greedy, and DQE is the valid core of the paper, and the SOTA claim should either be removed or backed by identical-base-model baseline runs.
minor comments (6)
  1. [§3.2] The description of K-Center-Greedy is imprecise: the statement that the algorithm 'iteratively selects the sample that is farthest from the current vector center' suggests a single center rather than the set of already selected points; the algorithm should be described as selecting the point with the largest distance to the nearest selected point.
  2. [§4.3] The phrase 'the second stage of DeepSpeed's data parallel training strategy' is ambiguous; it should say 'ZeRO stage 2' to be clear about the distributed training configuration.
  3. [Table 3] The p-value for SST-2 is formatted inconsistently as '< 0.0001 (6.14e-05)'; either report the exact p-value or use a uniform format across all rows.
  4. [Figure 4] Several example texts in Figure 4 contain corrupted glyphs (for example, '濠濥' and similar artifacts) that make the examples difficult or impossible to read; clean English text should be used.
  5. [§5.2] The claim that many test-set errors are actually correct predictions on mislabeled test examples is supported by only a few selected examples; a systematic random-sample annotation with counts and inter-annotator agreement would be needed to make this point quantitative.
  6. [General] The paper would benefit from a statement on the availability of code and the exact selected data subsets; without these, the DQE procedure is not fully reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed half-data advantage is an empirical test-set result, not a fitted or self-referential construction.

full rationale

The paper's derivation chain is not circular in any of the enumerated senses. DQE selects half the training data by K-Center-Greedy, fine-tunes Qwen2-7B-Instruct on that sampled subset, and then labels incorrectly predicted unsampled examples as uncovered, difficult, or noisy on the basis of nearest-neighbor label agreement. The final DQE training set is a mixture of the greedy sample, selected errors, and GPT-4o-verified removals, but the reported accuracy in Table 2 is measured on held-out public test sets (MR, CR, IMDb, SST-2, SST-5, AG News) that are not used to set any parameter or threshold of the method. No equation or definition reduces the test accuracy to the selection rule by construction: the rule is a heuristic data-selection hypothesis, and the evaluation is a genuinely external empirical check. The assumptions that similar texts usually share a label and that GPT-4o can verify label noise are substantive empirical premises, not definitions equivalent to the outcome, so any concern about them belongs to correctness risk rather than circularity. There is no load-bearing self-citation, no imported uniqueness theorem, and no fitted parameter renamed as a prediction. The paper is self-contained against public benchmarks, which is the normal non-circular case.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The method's core mechanism assumes that closest neighbors in embedding space share labels and that GPT-4o can reliably judge label correctness. The free parameters are the sampling fraction K = 0.5 and the top-1 similarity rule with no threshold. No new physical or mathematical entities are introduced.

free parameters (2)
  • sampling_fraction_k = 0.5
    K is set to half the preprocessed training set by hand as a 'balanced' choice; no sensitivity analysis is provided.
  • similarity_threshold = none (top-1)
    The method uses the top-1 nearest neighbor without a similarity threshold; the paper acknowledges a threshold could be set but does not explore it.
assumptions (3)
  • domain assumption Two highly similar data items have the same label, so a nearest neighbor with a different label indicates label noise.
    Invoked in Section 3.3 to categorize incorrectly predicted samples as noisy.
  • domain assumption Incorrectly predicted samples whose nearest neighbor lies in the unsampled set are 'uncovered' and those whose nearest neighbor lies in the sampled set are 'difficult'.
    Invoked in Section 3.3 to justify adding these samples to the training set.
  • domain assumption GPT-4o's judgments of label correctness are reliable enough to delete or retain samples.
    Used as the final arbiter in noisy-data detection (Section 3.3 and 5.2).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data Quality Enhancement on the Basis of Diversity with Large Language Models for Text Classification: Uncovered, Difficult, and Noisy." pith.science (2026). https://pith.science/paper/I76FJTUE

@misc{pith2026241206575,
  author       = {Pith},
  title        = {Pith review of: Data Quality Enhancement on the Basis of Diversity with Large Language Models for Text Classification: Uncovered, Difficult, and Noisy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I76FJTUE}},
  note         = {Machine review of arXiv:2412.06575}
}
read the original abstract

In recent years, the use of large language models (LLMs) for text classification has attracted widespread attention. Despite this, the classification accuracy of LLMs has not yet universally surpassed that of smaller models. LLMs can enhance their performance in text classification through fine-tuning. However, existing data quality research based on LLMs is challenging to apply directly to solve text classification problems. To further improve the performance of LLMs in classification tasks, this paper proposes a data quality enhancement (DQE) method for text classification based on LLMs. This method starts by using a greedy algorithm to select data, dividing the dataset into sampled and unsampled subsets, and then performing fine-tuning of the LLMs using the sampled data. Subsequently, this model is used to predict the outcomes for the unsampled data, categorizing incorrectly predicted data into uncovered, difficult, and noisy data. Experimental results demonstrate that our method effectively enhances the performance of LLMs in text classification tasks and significantly improves training efficiency, saving nearly half of the training time. Our method has achieved state-of-the-art performance in several open-source classification tasks.

Figures

Figures reproduced from arXiv: 2412.06575 by the authors.

Figure 1
Figure 1. The overall structure of the DQE method. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The identification process of Uncovered, Difficulty, and Noisy. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. the proportion between uncovered, difficult, [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Example of noisy data found by the DQE. 5.2 Noisy Data Experimental results indicate that even widely used public datasets commonly contain noisy data, es￾pecially within larger datasets where the issue of noisy data is more pronounced [PITH_FULL_IMAGE:figures/full_fi…
Figure 5
Figure 5. Figure 5: Examples of prediction results of DQE on the test set. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 7 canonical work pages

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. https://arxiv.org/pdf/2303.08774 Gpt-4 technical report . arXiv preprint arXiv:2303.08774

  2. [2]

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. 2023. https://arxiv.org/pdf/2309.16609 Qwen technical report . arXiv preprint arXiv:2309.16609

  3. [3]

    Martin Juan Jos \'e Bucher and Marco Martini. 2024. https://arxiv.org/pdf/2406.08660 Fine-tuned'small'llms (still) significantly outperform zero-shot generative ai models in text classification . arXiv preprint arXiv:2406.08660

  4. [4]

    Jacob Devlin. 2018. https://aclanthology.org/N19-1423.pdf Bert: Pre-training of deep bidirectional transformers for language understanding . arXiv preprint arXiv:1810.04805

  5. [5]

    Qianlong Du, Chengqing Zong, and Jiajun Zhang. 2023. Mods: Model-oriented data selection for instruction tuning. arXiv preprint arXiv:2311.15653

  6. [6]

    Zhengxiao Du, Yujie Qian, Xiao Liu, Ming Ding, Jiezhong Qiu, Zhilin Yang, and Jie Tang. 2021. https://arxiv.org/pdf/2103.10360 Glm: General language model pretraining with autoregressive blank infilling . arXiv preprint arXiv:2103.10360

  7. [7]

    Aleksandra Edwards and Jose Camacho-Collados. 2024. https://arxiv.org/pdf/2403.17661 Language models for text classification: Is in-context learning enough? arXiv preprint arXiv:2403.17661

  8. [8]

    Yuan Ge, Yilun Liu, Chi Hu, Weibin Meng, Shimin Tao, Xiaofeng Zhao, Hongxia Ma, Li Zhang, Hao Yang, and Tong Xiao. 2024. Clustering and Ranking: Diversity-preserved Instruction Selection through Expert-aligned Quality Estimation Clustering and ranking: Diversity-preserved instruction selection through expert-aligned quality estimation . arXiv preprint arX...

Show all 37 references
  1. [9]

    Minqing Hu and Bing Liu. 2004. https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=299a26b7c84b532878eb4b4ff45003042f6f9423 Mining and summarizing customer reviews . In Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mini...

  2. [10]

    Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei. 2020. https://arxiv.org/pdf/2001.08361 Scaling laws for neural language models . arXiv preprint arXiv:2001.08361

  3. [11]

    Kamran Kowsari, Kiana Jafari Meimandi, Mojtaba Heidarysafa, Sanjana Mendu, Laura Barnes, and Donald Brown. 2019. https://www.mdpi.com/2078-2489/10/4/150?source=post_page--------------------------- Text classification algorithms: A survey . Information, 10(4):150

  4. [12]

    Ming Li, Lichang Chen, Jiuhai Chen, Shwai He, and Tianyi Zhou. 2023 a . https://openreview.net/forum?id=xaqoZZqkPU Reflection-tuning: Recycling data for better instruction-tuning . In NeurIPS 2023 Workshop on Instruction Tuning and Instruction Following

  5. [13]

    Ming Li, Yong Zhang, Zhitao Li, Jiuhai Chen, Lichang Chen, Ning Cheng, Jianzong Wang, Tianyi Zhou, and Jing Xiao. 2024 a . https://aclanthology.org/2024.naacl-long.421 From quantity to quality: Boosting LLM performance with self-guided data selection for instruction tuning . I...

  6. [14]

    Qian Li, Hao Peng, Jianxin Li, Congying Xia, Renyu Yang, Lichao Sun, Philip S Yu, and Lifang He. 2022. A survey on text classification: From traditional to deep learning. ACM Transactions on Intelligent Systems and Technology (TIST), 13(2):1--41

  7. [15]

    Yinheng Li, Rogerio Bonatti, Sara Abdali, Justin Wagle, and Kazuhito Koishida. 2024 b . https://arxiv.org/pdf/2407.12813 Data generation using large language models for text classification: An empirical case study . arXiv preprint arXiv:2407.12813

  8. [16]

    Yuanzhi Li, S \'e bastien Bubeck, Ronen Eldan, Allie Del Giorno, Suriya Gunasekar, and Yin Tat Lee. 2023 b . https://arxiv.org/pdf/2309.05463 Textbooks are all you need ii: phi-1.5 technical report . arXiv preprint arXiv:2309.05463

  9. [17]

    Zhuoyan Li, Hangxiao Zhu, Zhuoran Lu, and Ming Yin. 2023 c . https://arxiv.org/pdf/2310.07849 Synthetic data generation with large language models for text classification: Potential and limitations . arXiv preprint arXiv:2310.07849

  10. [18]

    Andrew Maas, Raymond E Daly, Peter T Pham, Dan Huang, Andrew Y Ng, and Christopher Potts. 2011. https://aclanthology.org/P11-1015/ Learning word vectors for sentiment analysis . In Proceedings of the 49th annual meeting of the association for computational linguistics: Human l...

  11. [19]

    Shervin Minaee, Nal Kalchbrenner, Erik Cambria, Narjes Nikzad, Meysam Chenaghlu, and Jianfeng Gao. 2021. Deep learning--based text classification: a comprehensive review. ACM computing surveys (CSUR), 54(3):1--40

  12. [20]

    Shervin Minaee, Tomas Mikolov, Narjes Nikzad, Meysam Chenaghlu, Richard Socher, Xavier Amatriain, and Jianfeng Gao. 2024. https://arxiv.org/pdf/2402.06196 Large language models: A survey . arXiv preprint arXiv:2402.06196

  13. [21]

    Bo Pang and Lillian Lee. 2005. https://arxiv.org/pdf/cs/0506075 Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales . arXiv preprint cs/0506075

  14. [22]

    Libo Qin, Qiguang Chen, Xiachong Feng, Yang Wu, Yongheng Zhang, Yinghui Li, Min Li, Wanxiang Che, and Philip S Yu. 2024. https://arxiv.org/pdf/2405.12819 Large language models meet nlp: A survey . arXiv preprint arXiv:2405.12819

  15. [23]

    Samyam Rajbhandari, Jeff Rasley, Olatunji Ruwase, and Yuxiong He. 2020. Zero: Memory optimizations toward training trillion parameter models. In SC20: International Conference for High Performance Computing, Networking, Storage and Analysis, pages 1--16. IEEE

  16. [24]

    Ozan Sener and Silvio Savarese. 2018. https://openreview.net/forum?id=H1aIuk-RW Active learning for convolutional neural networks: A core-set approach . In International Conference on Learning Representations

  17. [25]

    Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, Christopher D Manning, Andrew Y Ng, and Christopher Potts. 2013. https://aclanthology.org/D13-1170.pdf Recursive deep models for semantic compositionality over a sentiment treebank . In Proceedings of the 2013 conference o...

  18. [26]

    Gemma Team, Thomas Mesnard, Cassidy Hardin, Robert Dadashi, Surya Bhupatiraju, Shreya Pathak, Laurent Sifre, Morgane Rivi \`e re, Mihir Sanjay Kale, Juliette Love, et al. 2024 a . https://arxiv.org/pdf/2403.08295 Gemma: Open models based on gemini research and technology . arX...

  19. [27]

    Gemma Team, Morgane Riviere, Shreya Pathak, Pier Giuseppe Sessa, Cassidy Hardin, Surya Bhupatiraju, L \'e onard Hussenot, Thomas Mesnard, Bobak Shahriari, Alexandre Ram \'e , et al. 2024 b . https://arxiv.org/pdf/2408.00118 Gemma 2: Improving open language models at a practica...

  20. [28]

    Kushal Tirumala, Daniel Simig, Armen Aghajanyan, and Ari Morcos. 2024. https://proceedings.neurips.cc/paper_files/paper/2023/file/a8f8cbd7f7a5fb2c837e578c75e5b615-Paper-Datasets_and_Benchmarks.pdf D4: Improving llm pretraining via document de-duplication and diversification . ...

  21. [29]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timoth \'e e Lacroix, Baptiste Rozi \`e re, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023 a . https://arxiv.org/pdf/2302.13971 Llama: Open and efficient foundation language models . arXiv...

  22. [30]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023 b . https://arxiv.org/pdf/2307.09288 Llama 2: Open foundation and fine-tuned chat models . arXiv preprint arX...

  23. [31]

    An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, et al. 2024. https://arxiv.org/pdf/2407.10671 Qwen2 technical report . arXiv preprint arXiv:2407.10671

  24. [32]

    Zhiyuan Zeng, Jiatong Yu, Tianyu Gao, Yu Meng, Tanya Goyal, and Danqi Chen. 2023. https://arxiv.org/pdf/2310.07641 Evaluating large language models at evaluating instruction following . arXiv preprint arXiv:2310.07641

  25. [33]

    Shengyu Zhang, Linfeng Dong, Xiaoya Li, Sen Zhang, Xiaofei Sun, Shuhe Wang, Jiwei Li, Runyi Hu, Tianwei Zhang, Fei Wu, et al. 2023. https://arxiv.org/pdf/2308.10792 Instruction tuning for large language models: A survey . arXiv preprint arXiv:2308.10792

  26. [34]

    Xiang Zhang, Junbo Zhao, and Yann LeCun. 2015. https://proceedings.neurips.cc/paper/2015/file/250cf8b51c773f3f8dc8b4be867a9a02-Paper.pdf Character-level convolutional networks for text classification . Advances in neural information processing systems, 28

  27. [35]

    Yazhou Zhang, Mengyao Wang, Chenyu Ren, Qiuchi Li, Prayag Tiwari, Benyou Wang, and Jing Qin. 2024. Pushing the limit of llm capacity for text classification. arXiv preprint arXiv:2402.07470

  28. [36]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  29. [37]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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