Pith. sign in

REVIEW 5 major objections 4 minor 45 references

CHAI for LLMs: Improving Code-Mixed Translation in Large Language Models through Reinforcement Learning with AI Feedback

T0 review · 5 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that CHAI, an RLAIF pipeline using GPT-4o preference labels, lifts human-preferred Hinglish translation quality of a Llama-3.1 model by 25.66% in win rate.

desk verdict Useful first RLAIF-for-code-mixed-translation recipe, but the 25.66% headline is statistically unsupported; the paper deserves review but only after the authors add significance testing and temper claims. read the letter →

arxiv 2411.09073 v3 pith:RH2RGGQ4 submitted 2024-11-13 cs.CL cs.AIcs.LG

Wenbo Zhang , Aditya Majumdar , Amulya Yadav This is my paper · ORCID
classification cs.CLcs.AIcs.LG
keywords code-mixedtranslationHinglishRLAIFLLMasannotatormachinepreferencedataPPOLlama-3.1-8B-Instruct
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 claims that a general-purpose alignment pipeline called CHAI — reinforcement learning from AI feedback (RLAIF) built on GPT-4o preference labels — can substantially improve how large language models translate monolingual English text into code-mixed Hinglish. The authors report that the CHAI-tuned Llama-3.1-8B-Instruct model beats state-of-the-art open baselines by 25.66% in human-adjudicated win rate on code-mixed translation, and also outscores them on COMET and on downstream Hinglish sentiment classification. The argument matters because human preference annotation for code-mixed languages is scarce and expensive; if LLM annotators can supply reliable preference data, alignment-based improvement becomes scalable for other low-resource code-mixed languages. The paper also reports a cautionary internal result: supervised fine-tuning on the parallel corpus hurt final quality, and chrF-style lexical metrics favor the SFT baseline even where human and LLM judges prefer CHAI outputs.

What carries the argument

The load-bearing machinery is the AI preference-labeling loop. CHAI pairs each English source sentence with two Hinglish translations, has GPT-4o (with randomly switched candidate order, three temperature settings, and majority voting) pick the better one, trains a reward model on those binary labels via the Bradley-Terry model, and optimizes the policy with PPO using a KL penalty to stay near the base distribution. The same preference pipeline is also used to compute win rates, with Gemini-1.5-Flash and three human annotators acting as judges. The paper's key finding about this machinery is that simple rule-augmented prompts outperform chain-of-thought and few-shot prompting for LLM annotators on code-mixed data, and that the SFT stage of a conventional RLAIF recipe is counterproductive in this setting.

What would settle it

Run pairwise human evaluation of CHAI-LLM versus the strongest SFT baseline on the MixMT and HinGE test sets with at least 20 native Hinglish speakers, record individual preferences, and compute a confidence interval for the win rate; if the interval contains 50% or excludes the reported margin, the central human-preference claim is not supported. An inter-annotator agreement no better than chance would similarly undermine the averaging.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that preference-based alignment works for code-mixed translation when the preference labels come from an AI annotator. CHAI first asks GPT-4o, at three temperatures, to choose the better of two Hinglish translations of the same English sentence; majority vote gives a binary label. Those labels train a Bradley-Terry reward model, and a PPO policy update with a KL penalty aligns Llama-3.1-8B-Instruct to the reward. The authors find that basic rule-augmented prompts align LLM annotators with human judges 61.8% of the time, and that the resulting CHAI-LLM wins human pairwise comparisons against the base model, two SFT baselines, and a DPO baseline, while improving Hinglish sentiment accuracy and showing mostly positive cross-lingual transfer to English-Bengali and English-Spanish. The lexical metrics chrF and chrF++ do not track the human preference result, which the authors attribute to known weaknesses of n-gram metrics for code-mixed output.

Load-bearing premise

The entire 25.66% human-preference claim rests on majority votes from only three convenience-recruited annotators, with no confidence intervals, significance tests, or inter-annotator agreement reported, and the human win rate over the strongest SFT baseline is only 52.15%.

Editorial extensions

If this is right

  • If the win-rate results hold, LLM-generated preference labels can replace expensive human annotation for code-mixed translation alignment, making RLAIF practical for other low-resource language pairs.
  • The CHAI-trained model should generalize beyond translation: the same alignment improved Hinglish sentiment accuracy and F1 by double digits on two datasets.
  • The cross-lingual gains, especially to English-Spanish code-mixing, suggest some preference knowledge learned on Hinglish transfers to other code-mixed pairs, while English-French did not clearly benefit.
  • Evaluation of code-mixed systems should weight human or LLM preference over chrF and chrF++, since the lexical metrics contradict the judged preference ordering.
  • Because SFT lowered final quality, instruction-tuned base models may need a modified RLAIF recipe that skips or rethinks the supervised fine-tuning stage.

Reading between the lines

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

  • Beyond the paper: because the LLM annotator agrees with humans only about 62% of the time, preference data carries an annotator-model-specific bias; a different labeler than GPT-4o could produce a differently flavored 'best Hinglish'.
  • Beyond the paper: the human evaluation's three-person convenience sample means the claimed 25.66% margin is a point estimate with unknown spread; rerunning with a larger, more diverse annotator pool would likely shrink the margin even if the direction survives.
  • Beyond the paper: the counterproductive SFT result suggests code-mixed corpus noise, not model capacity, may be the binding constraint; a testable extension is to filter noisy parallel data before SFT and see whether quality improves.
  • Beyond the paper: the same RLAIF recipe could be applied to code-mixed summarization or language identification once preference data are generated, since the sentiment gains suggest task transfer within code-mixed understanding.
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

5 major / 4 minor

Summary. The paper proposes CHAI, a reinforcement-learning-from-AI-feedback (RLAIF) framework for improving code-mixed machine translation, specifically English-to-Hinglish. The pipeline uses GPT-4o to annotate pairwise preferences among candidate translations, trains a Bradley-Terry reward model on those AI labels, and fine-tunes Llama-3.1-8B-Instruct with PPO. The central claim is that CHAI-powered LLMs outperform state-of-the-art open-source LLMs by 25.66% in human-adjudicated win rate on code-mixed translation, with additional evidence from automatic metrics, cross-lingual transfer experiments, and code-mixed sentiment analysis.

Significance. If the headline result is correct, the paper would be a valuable first demonstration of RLAIF for code-mixed NLP, with a practical recipe (LLM-as-annotator, reward modeling, PPO) that could extend to other low-resource code-mixed tasks. The authors contribute useful ablations: the comparison with DPO (Table A5), the SFT-on/off study (Table 2), and the temperature sweep (Figure 2). The code is promised to be public. However, the contribution's weight rests almost entirely on the human-preference evaluation, and that evaluation is not yet statistically grounded. The cross-lingual results also contain an internal contradiction with the paper's own automatic metrics.

major comments (5)
  1. [Section 4, 'Evaluation Metrics' and Figure 3; Appendix A.14] The headline claim of a 25.66% win-rate improvement is not supported by statistically meaningful evidence. Against the stronger SFT baseline πsft-1, the human-adjudicated win rate is only 52.15% (Figure 3). With the reported test sizes (376 MixMT + 395 HinGE sentences, possibly averaged across these), a 52.15% win rate is not distinguishable from 50% (approximate 95% CI roughly 48.6–55.7%, p > 0.2). No confidence intervals, significance tests, or inter-annotator agreement are reported anywhere in Section 4 or Appendix A.14. Because the automatic metrics in Table 3 are mixed — CHAI is below the SFT baselines on chrF/chrF++ on both datasets — the human preference result is the sole support for the abstract's claim, and as reported it does not establish a robust advantage.
  2. [Abstract and Section 4, Figure 3] The paper never defines how the 25.66% figure is computed. Figure 3 reports per-baseline win rates (60.65%, 52.15%, 76.91% against πbase, πsft-1, πsft-2, respectively), but the abstract and introduction quote 25.66% without a formula or a description of the aggregation. The manuscript should state exactly how this number is derived (e.g., average of win-rate margins above 50%) and should report the per-comparison win rates with uncertainty bounds for each dataset separately.
  3. [Section 4, 'Cross-lingual Transferability' and Table 4] The text claims that 'all classic metrics consistently favor CHAI-LLM' when describing cross-lingual transfer, but Table 4 contradicts this for the English→French direction: chrF drops from 34.88 (πbase) to 22.07 (CHAI-LLM), and chrF++ drops from 31.52 to 19.85. The later statement that the RLAIF procedure improves transfer on 'two out of three language pairs' is accurate for chrF, but the 'consistently favor' sentence is incorrect and must be corrected. The near-chance LLM win rate for English→Bengali (50.11%) also undercuts the strength of the cross-lingual claim.
  4. [Table 3 and Section 4, 'Impact of CHAI on Translation Quality'] The automatic metrics do not corroborate the human-preference claim. On both MixMT 2022 and HinGE, CHAI-LLM has lower chrF and chrF++ than both SFT baselines (e.g., MixMT chrF: 42.68 vs. 48.57 for πsft-1), while COMET increases by only about 0.02–0.04. The paper dismisses this with a general reference to Winata et al. (2024), but no evidence is provided that chrF is an invalid indicator for Hinglish code-mixed translation specifically. The authors should either report metric–human correlation on this task or discuss more directly why the SFT baselines are preferred by chrF while the human evaluators (weakly) prefer CHAI.
  5. [Table 1 and Section 3, 'Collecting Preference Data Using LLMs'] A core contribution is the claim that LLM annotators can accurately label code-mixed translation preferences. The maximum alignment with human preferences reported in Table 1 is 61.8%, which is modest and is measured on a single 1000-sample evaluation set (Appendix A.3) with no confidence intervals or inter-annotator agreement for the human labels. The reward model is trained on these noisy AI labels, so the entire RLAIF pipeline is only as trustworthy as the 61.8% alignment. The paper should provide a more complete assessment of this agreement (per-category, with variance) and discuss the implications of the noise for reward-model training.
minor comments (4)
  1. [Section 4, 'Evaluation Metrics'] The description of the human evaluation protocol mentions only the MixMT 2022 test set, while Figure 3 and the text refer to win rates on both MixMT 2022 and HinGE. Please clarify which test sets were used for human adjudication and whether the reported win rates are per-dataset or averaged.
  2. [Abstract and Conclusion] The phrase 'outperform ... by 25.66% (in terms of win rate)' is ambiguous because win rate is bounded by 0–100%, and the reader cannot tell whether 25.66% is an absolute margin over 50%, a relative improvement, or an average of different margins. Please define the quantity precisely.
  3. [Throughout] There are several typographical errors and inconsistent spellings: 'developmennt' (Section 7), 'renumeration' (Section 7), 'alternated' in the human-evaluation rules (Appendix A.10), 'RAILF' instead of 'RLAIF' (Appendix A.5), and inconsistent 'Llama'/'LlaMA' spellings. A thorough proofread is needed.
  4. [Appendix A.14] The recruitment details state only that three annotators were recruited from a university via convenience sampling and paid $25/hour; no information is given about proficiency screening, annotator agreement, or the decision to use only three annotators. This is material to the reliability of the headline result.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity; the CHAI claim rests on external human and Gemini adjudication, not on the fitted reward model.

full rationale

The paper's derivation chain is not circular in the sense defined by the review criteria. The reward model is trained on GPT-4o preference labels, while the headline 25.66% win-rate claim is adjudicated by three human annotators on held-out test sets (and, for the LLM-evaluation win rates, by Gemini-1.5-Flash). The final evaluation is therefore not statistically forced by the fitted reward model or by the AI feedback used in training. No equation in the paper reduces the reported win rate to the training objective or to the preference-labeling procedure: Equation (1)-(3) define reward-model and RL training, whereas the win rate in Section 4 is computed from independent preference judgments on CHAI versus baseline outputs. The selection of the best annotation prompt in Table 1 does use human labels, but those labels are on a 1000-point subsample and serve only to choose a prompting strategy; the final human evaluation is on separate MixMT 2022 and HinGE test sets, so the central claim is not a fitted input renamed as a prediction. There is one self-citation (Zhang et al., 2023c, cited for prior pre-training-based code-mixed augmentation attempts), but it is not load-bearing: the CHAI method and its evaluation do not depend on that citation's validity. The more serious concerns in this paper are statistical robustness issues rather than circularity: only three convenience-sampled human annotators were used, no confidence intervals or significance tests are reported, and the human win rate against the stronger SFT baseline is 52.15%, which is near chance. Those concerns affect the strength of the empirical claim, not whether the claim is derived from its own inputs.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim relies on two empirical proxies: GPT-4o's preference labels (61.8% agreement with humans on a 1000-example sample) and a three-annotator human gold standard. The RL machinery is standard. No new entities are introduced. The main free parameters are the KL coefficient and decoding temperature, both tuned on the evaluation test sets.

free parameters (4)
  • KL divergence coefficient eta = 0.04
    Selected via ablation study (A.7) to maximize final model performance on the evaluation test sets; a test-set-tuned hyperparameter.
  • Decoding temperature T = 0.6
    Chosen from Figure 2, which varies temperature and reports chrF, chrF++, COMET; optimal at T=0.6 on the evaluation test sets.
  • LLM annotator temperature set = {0.1, 0.3, 0.5} with majority vote
    Chosen for preference labeling; no ablation shown for the temperature set itself.
  • Prompt strategy for annotation = basic + rule zero-shot
    Selected as the best prompt from Table 1 based on alignment with a 1000-example human-labeled set; a design choice with a reported 61.8% alignment.
assumptions (6)
  • standard math Bradley-Terry model of pairwise preferences
    Used in Eq. 1 to define the reward model loss; standard statistical model.
  • standard math PPO is a valid policy optimization algorithm
    Used in Stage 3; standard RL algorithm.
  • domain assumption GPT-4o preference labels are a sufficient proxy for human preferences in code-mixed translation
    The entire Stage 2 depends on this; measured alignment is only 61.8% on a 1000-example sample (Table 1), which is moderate.
  • domain assumption The human evaluators' majority vote is a reliable gold standard
    Only three annotators, convenience-sampled from a university (A.14), no inter-annotator agreement reported.
  • domain assumption The MixMT, HinGE, and MT-Aug test sets are representative of code-mixed translation quality
    Used for all main evaluations; small sizes (376-610 sentences).
  • domain assumption Instruction-tuned Llama-3.1-8B is an appropriate base model
    The framework is only tested on this model; results may not transfer.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CHAI for LLMs: Improving Code-Mixed Translation in Large Language Models through Reinforcement Learning with AI Feedback." pith.science (2026). https://pith.science/paper/RH2RGGQ4

@misc{pith2026241109073,
  author       = {Pith},
  title        = {Pith review of: CHAI for LLMs: Improving Code-Mixed Translation in Large Language Models through Reinforcement Learning with AI Feedback},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RH2RGGQ4}},
  note         = {Machine review of arXiv:2411.09073}
}
read the original abstract

Large Language Models (LLMs) have demonstrated remarkable capabilities across various NLP tasks but struggle with code-mixed (or code-switched) language understanding. For example, prior work benchmarking the performance of multilingual LLMs on code-mixed translation tasks has demonstrated that current state-of-the-art multilingual LLMs are ineffective in dealing with code-mixed languages. However, the question of how to improve the capability of multilingual LLMs to handle code-mixed language has not received any attention to date. In this paper, we tackle this research gap by proposing CHAI, a novel general-purpose framework for improving the ability of multilingual LLMs to handle code-mixed languages. CHAI relies on three novel contributions made in this paper. First, we explore the ability of LLMs to provide accurate annotations for code-mixed translation tasks. Second, we leverage this ability of LLMs as annotators to generate preference data for code-mixed translation tasks at scale, which are then used within a reinforcement learning from AI feedback (RLAIF) procedure to improve LLMs' capability on code-mixed tasks. Third, we conduct a rigorous experimental evaluation across various real-world datasets and settings. Our analysis shows that CHAI-powered LLMs outperform state-of-the-art open-source LLMs by 25.66% (in terms of win rate adjudicated by human annotators) in code-mixed translation tasks. This work represents a first step towards developing more inclusive code-mixed LLMs.

Figures

Figures reproduced from arXiv: 2411.09073 by the authors.

Figure 1
Figure 1. Overall architecture of the RLAIF Procedure used in CHAI. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Relationship between the temperature and the [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Corresponding win rate to measure CHAI’s ability in improving code mixed translation ability. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 12 canonical work pages

  1. [1]

    Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, et al. 2022. Constitutional ai: Harmlessness from ai feedback. arXiv preprint arXiv:2212.08073

    arXiv 2022

  2. [2]

    Ralph Allan Bradley and Milton E Terry. 1952. Rank analysis of incomplete block designs: I. the method of paired comparisons. Biometrika, 39(3/4):324--345

    1952

  3. [3]

    Szu-Wei Cheng, Chung-Wen Chang, Wan-Jung Chang, Hao-Wei Wang, Chih-Sung Liang, Taishiro Kishimoto, Jane Pei-Chen Chang, John S Kuo, and Kuan-Pin Su. 2023. The now and future of chatgpt and gpt in psychiatry. Psychiatry and clinical neurosciences, 77(11):592--596

    work page 2023

  4. [4]

    A Conneau. 2019. Unsupervised cross-lingual representation learning at scale. arXiv preprint arXiv:1911.02116

    arXiv 2019

  5. [5]

    Mrinal Dhar, Vaibhav Kumar, and Manish Shrivastava. 2018. https://aclanthology.org/W18-3817/ Enabling code-mixed translation: Parallel corpus creation and MT augmentation approach . In Proceedings of the First Workshop on Linguistic Resources for Natural Language Processing, pages 131--140, Santa Fe, New Mexico, USA. Association for Computational Linguistics

    work page 2018

  6. [6]

    Aaron Grattafiori and et. al. 2024. https://arxiv.org/abs/2407.21783 The llama 3 herd of models . Preprint, arXiv:2407.21783

    arXiv 2024

  7. [8]

    Ayushman Gupta, Akhil Bhogal, and Kripabandhu Ghosh. 2024 b . Code-mixer ya nahi: Novel approaches to measuring multilingual llms' code-mixing capabilities. arXiv preprint arXiv:2410.11079

    arXiv 2024

  8. [9]

    Ayushman Gupta, Akhil Bhogal, and Kripabandhu Ghosh. 2024 c . Multilingual controlled generation and gold-standard-agnostic evaluation of code-mixed sentences. arXiv preprint arXiv:2410.10580

    arXiv 2024

Show all 45 references
  1. [10]

    Zhiwei He, Xing Wang, Wenxiang Jiao, Zhuosheng Zhang, Rui Wang, Shuming Shi, and Zhaopeng Tu. 2024. https://doi.org/10.18653/v1/2024.naacl-long.451 Improving machine translation with human feedback: An exploration of quality estimation as a reward model . In Proceedings of the...

    2024 doi

  2. [11]

    Krystal Hu. 2023. https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/ Chatgpt sets record for fastest-growing user base - analyst note

    2023

  3. [12]

    Natasha Jaques, Asma Ghandeharioun, Judy Hanwen Shen, Craig Ferguson, Agata Lapedriza, Noah Jones, Shixiang Gu, and Rosalind Picard. 2019. Way off-policy batch deep reinforcement learning of implicit human preferences in dialog. arXiv preprint arXiv:1907.00456

    2019 arXiv

  4. [13]

    Yuhang Lai, Siyuan Wang, Shujun Liu, Xuanjing Huang, and Zhongyu Wei. 2024. https://doi.org/10.18653/v1/2024.findings-acl.465 AL a RM : Align language models via hierarchical rewards modeling . In Findings of the Association for Computational Linguistics: ACL 2024, pages 7817-...

    2024 doi

  5. [14]

    Cassidy Laidlaw, Shivam Singhal, and Anca Dragan. 2024. Preventing reward hacking with occupancy measure regularization. arXiv preprint arXiv:2403.03185

    2024 arXiv

  6. [15]

    Guillaume Lample and Alexis Conneau. 2019. Cross-lingual language model pretraining. arXiv preprint arXiv:1901.07291

    2019 arXiv

  7. [16]

    Harrison Lee, Samrat Phatale, Hassan Mansoor, Thomas Mesnard, Johan Ferret, Kellie Lu, Colton Bishop, Ethan Hall, Victor Carbune, Abhinav Rastogi, and Sushant Prakash. 2024. https://arxiv.org/abs/2309.00267 Rlaif vs. rlhf: Scaling reinforcement learning from human feedback wit...

    2024 arXiv

  8. [17]

    Ang Li, Qiugen Xiao, Peng Cao, Jian Tang, Yi Yuan, Zijie Zhao, Xiaoyuan Chen, Liang Zhang, Xiangyang Li, Kaitong Yang, Weidong Guo, Yukang Gan, Xu Yu, Daniell Wang, and Ying Shan. 2024. https://arxiv.org/abs/2403.08309 Hrlaif: Improvements in helpfulness and harmlessness in op...

    2024 arXiv

  9. [18]

    Alexandre Magueresse, Vincent Carles, and Evan Heetderks. 2020. Low-resource languages: A review of past work and future challenges. arXiv preprint arXiv:2006.07264

    2020 arXiv

  10. [19]

    OpenAI, Josh Achiam, and et. al. 2024. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . Preprint, arXiv:2303.08774

    2024 arXiv

  11. [20]

    Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and...

    2022 arXiv

  12. [21]

    Yang Jeong Park, Daniel Kaplan, Zhichu Ren, Chia-Wei Hsu, Changhao Li, Haowei Xu, Sipei Li, and Ju Li. 2024. Can chatgpt be used to generate scientific hypotheses? Journal of Materiomics, 10(3):578--584

    2024

  13. [22]

    Parth Patwa, Gustavo Aguilar, Sudipta Kar, Suraj Pandey, Srinivas PYKL, Björn Gambäck, Tanmoy Chakraborty, Thamar Solorio, and Amitava Das. 2020. https://arxiv.org/abs/2008.04277 Semeval-2020 task 9: Overview of sentiment analysis of code-mixed tweets . Preprint, arXiv:2008.04277

    2020 arXiv

  14. [23]

    Pouya Pezeshkpour and Estevam Hruschka. 2023. Large language models sensitivity to the order of options in multiple-choice questions. arXiv preprint arXiv:2308.11483

    2023 arXiv

  15. [24]

    Maja Popovi \'c . 2015. chrf: character n-gram f-score for automatic mt evaluation. In Proceedings of the tenth workshop on statistical machine translation, pages 392--395

    2015

  16. [25]

    Maja Popovi \'c . 2017. chrf++: words helping character n-grams. In Proceedings of the second conference on machine translation, pages 612--618

    2017

  17. [26]

    Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9

    2019

  18. [27]

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2023. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36:53728--53741

    2023

  19. [28]

    Md Nishat Raihan, Dhiman Goswami, Antara Mahmud, Antonios Anstasopoulos, and Marcos Zampieri. 2023. Sentmix-3l: A bangla-english-hindi code-mixed dataset for sentiment analysis. arXiv preprint arXiv:2310.18023

    2023 arXiv

  20. [29]

    Muhammad Ramzan, Aamir Aziz, and Maimoona Ghaffar. 2021. A study of code-mixing and code-switching (urdu and punjabi) in children's early speech. Journal of Language and Linguistic Studies, 17(2):869--881

    2021

  21. [30]

    Ricardo Rei, Craig Stewart, Ana C Farinha, and Alon Lavie. 2020. Comet: A neural framework for mt evaluation. arXiv preprint arXiv:2009.09025

    2020 arXiv

  22. [31]

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. 2017. https://arxiv.org/abs/1707.06347 Proximal policy optimization algorithms . Preprint, arXiv:1707.06347

    2017 arXiv

  23. [32]

    Vivek Srivastava and Mayank Singh. 2021. Hinge: A dataset for generation and evaluation of code-mixed hinglish text. arXiv preprint arXiv:2107.03760

    2021 arXiv

  24. [33]

    Vivek Srivastava and Mayank Singh. 2022. https://aclanthology.org/2022.wmt-1.74 Overview and results of M ix MT shared-task at WMT 2022 . In Proceedings of the Seventh Conference on Machine Translation (WMT), pages 806--811, Abu Dhabi, United Arab Emirates (Hybrid). Associatio...

    2022

  25. [34]

    Ishan Tarunesh, Syamantak Kumar, and Preethi Jyothi. 2021. From machine translation to code-switching: Generating high-quality code-switched text. arXiv preprint arXiv:2107.06483

    2021 arXiv

  26. [35]

    Gemini Team, Petko Georgiev, and et. al. 2024. https://arxiv.org/abs/2403.05530 Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context . Preprint, arXiv:2403.05530

    2024 arXiv

  27. [36]

    S Thara and Prabaharan Poornachandran. 2018. Code-mixing: A brief survey. In 2018 International conference on advances in computing, communications and informatics (ICACCI), pages 2382--2388. IEEE

    2018

  28. [37]

    Haoye Tian, Weiqi Lu, Tsz On Li, Xunzhu Tang, Shing-Chi Cheung, Jacques Klein, and Tegawend \'e F Bissyand \'e . 2023. Is chatgpt the ultimate programming assistant--how far is it? arXiv preprint arXiv:2304.11938

    2023 arXiv

  29. [38]

    Binghai Wang, Rui Zheng, Lu Chen, Yan Liu, Shihan Dou, Caishuang Huang, Wei Shen, Senjie Jin, Enyu Zhou, Chenyu Shi, et al. 2024. Secrets of rlhf in large language models part ii: Reward modeling. arXiv preprint arXiv:2401.06080

    2024 arXiv

  30. [39]

    Genta Indra Winata, David Anugraha, Lucky Susanto, Garry Kuwanto, and Derry Tanti Wijaya. 2024. Metametrics: Calibrating metrics for generation tasks using human preferences. arXiv preprint arXiv:2410.02381

    2024 arXiv

  31. [40]

    Haoran Xu, Young Jin Kim, Amr Sharaf, and Hany Hassan Awadalla. 2023. A paradigm shift in machine translation: Boosting translation performance of large language models. arXiv preprint arXiv:2309.11674

    2023 arXiv

  32. [41]

    Nuo Xu, Jun Zhao, Can Zu, Sixian Li, Lu Chen, Zhihao Zhang, Rui Zheng, Shihan Dou, Wenjuan Qin, Tao Gui, Qi Zhang, and Xuanjing Huang. 2024. https://arxiv.org/abs/2402.11525 Advancing translation preference modeling with rlhf: A step towards cost-effective solution . Preprint,...

    2024 arXiv

  33. [42]

    Ruochen Zhang, Samuel Cahyawijaya, Jan Christian Blaise Cruz, and Alham Fikri Aji. 2023 a . Multilingual large language models are not (yet) code-switchers. arXiv preprint arXiv:2305.14235

    2023 arXiv

  34. [43]

    Ruochen Zhang, Samuel Cahyawijaya, Jan Christian Blaise Cruz, Genta Winata, and Alham Fikri Aji. 2023 b . https://doi.org/10.18653/v1/2023.emnlp-main.774 Multilingual large language models are not (yet) code-switchers . In Proceedings of the 2023 Conference on Empirical Method...

    2023 doi

  35. [44]

    Wenbo Zhang, Hangzhi Guo, Prerna Ranganathan, Jay Patel, Sathyanath Rajasekharan, Nidhi Danayak, Manan Gupta, and Amulya Yadav. 2023 c . https://doi.org/10.1609/aaai.v37i12.26709 A continual pre-training approach to tele-triaging pregnant women in kenya . Proceedings of the AA...

    2023 doi

  36. [45]

    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...

  37. [46]

    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 12, 2026 · model on record in the stance chip above.