@yuchenlin\n\t congrats on this work!

Would you be able to link the dataset (https://huggingface.co/datasets/goodbadgreedy/GoodBadGreedy) to this paper?

See here on how to do that: https://huggingface.co/docs/hub/en/datasets-cards#linking-a-paper

\n","updatedAt":"2024-07-16T14:40:10.262Z","author":{"_id":"5f1158120c833276f61f1a84","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1608042047613-5f1158120c833276f61f1a84.jpeg","fullname":"Niels Rogge","name":"nielsr","type":"user","isPro":false,"isHf":true,"isHfAdmin":false,"isMod":false,"followerCount":1096,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7975223660469055},"editors":["nielsr"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1608042047613-5f1158120c833276f61f1a84.jpeg"],"reactions":[],"isReport":false,"parentCommentId":"6695cf90490f463859d6a39c"}}]},{"id":"6695fb6e080d0f34def6eb33","author":{"_id":"607f666a4ad99100d63ce35c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png","fullname":"Bill Yuchen Lin","name":"yuchenlin","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":145,"isUserFollowing":false},"createdAt":"2024-07-16T04:47:42.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"goodbadgreedy/GoodBadGreedy","html":"

goodbadgreedy/GoodBadGreedy

\n","updatedAt":"2024-07-16T04:47:42.392Z","author":{"_id":"607f666a4ad99100d63ce35c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png","fullname":"Bill Yuchen Lin","name":"yuchenlin","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":145,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5133254528045654},"editors":["yuchenlin"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png"],"reactions":[],"isReport":false}},{"id":"66971f299ad36faed14a957f","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false},"createdAt":"2024-07-17T01:32:25.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [UBENCH: Benchmarking Uncertainty in Large Language Models with Multiple Choice Questions](https://huggingface.co/papers/2406.12784) (2024)\n* [WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild](https://huggingface.co/papers/2406.04770) (2024)\n* [Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges](https://huggingface.co/papers/2406.12624) (2024)\n* [RUPBench: Benchmarking Reasoning Under Perturbations for Robustness Evaluation in Large Language Models](https://huggingface.co/papers/2406.11020) (2024)\n* [MHPP: Exploring the Capabilities and Limitations of Language Models Beyond Basic Code Generation](https://huggingface.co/papers/2405.11430) (2024)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2024-07-17T01:32:25.224Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.759681224822998},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2407.10457","authors":[{"_id":"6695cf864c450902c8cf3e5a","user":{"_id":"61d595f0bfc2372aaf42b766","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1675437310060-61d595f0bfc2372aaf42b766.jpeg","isPro":false,"fullname":"Yifan Song","user":"Solaris99","type":"user"},"name":"Yifan Song","status":"claimed_verified","statusLastChangedAt":"2025-05-13T09:54:35.762Z","hidden":false},{"_id":"6695cf864c450902c8cf3e5b","user":{"_id":"6490d4ba1afdee3acd1147f6","avatarUrl":"/avatars/ae13c7b21fe9ced7541dcd664d1b94ed.svg","isPro":false,"fullname":"Guoyin Wang","user":"guoyinwang","type":"user"},"name":"Guoyin Wang","status":"claimed_verified","statusLastChangedAt":"2024-12-03T11:17:06.403Z","hidden":false},{"_id":"6695cf864c450902c8cf3e5c","name":"Sujian Li","hidden":false},{"_id":"6695cf864c450902c8cf3e5d","user":{"_id":"607f666a4ad99100d63ce35c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png","isPro":false,"fullname":"Bill Yuchen Lin","user":"yuchenlin","type":"user"},"name":"Bill Yuchen Lin","status":"extracted_pending","statusLastChangedAt":"2024-07-16T01:40:23.582Z","hidden":false}],"publishedAt":"2024-07-15T06:12:17.000Z","submittedOnDailyAt":"2024-07-16T00:10:32.923Z","title":"The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore\n Non-Determinism","submittedOnDailyBy":{"_id":"607f666a4ad99100d63ce35c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png","isPro":false,"fullname":"Bill Yuchen Lin","user":"yuchenlin","type":"user"},"summary":"Current evaluations of large language models (LLMs) often overlook\nnon-determinism, typically focusing on a single output per example. This limits\nour understanding of LLM performance variability in real-world applications.\nOur study addresses this issue by exploring key questions about the performance\ndifferences between greedy decoding and sampling, identifying benchmarks'\nconsistency regarding non-determinism, and examining unique model behaviors.\nThrough extensive experiments, we observe that greedy decoding generally\noutperforms sampling methods for most evaluated tasks. We also observe\nconsistent performance across different LLM sizes and alignment methods, noting\nthat alignment can reduce sampling variance. Moreover, our best-of-N sampling\napproach demonstrates that smaller LLMs can match or surpass larger models such\nas GPT-4-Turbo, highlighting the untapped potential of smaller LLMs. This\nresearch shows the importance of considering non-determinism in LLM evaluations\nand provides insights for future LLM development and evaluation.","upvotes":24,"discussionId":"6695cf874c450902c8cf3e7d","githubRepo":"https://github.com/yifan-song793/goodbadgreedy","githubRepoAddedBy":"auto","ai_summary":"Research explores non-determinism in LLM evaluations, comparing greedy decoding and sampling, and finds that smaller models can match larger ones with best-of-N sampling.","ai_keywords":["greedy decoding","sampling","benchmarks","alignment methods","best-of-N sampling","LLM","language models"],"githubStars":30},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"668cd4bbe990292e5f6974d3","avatarUrl":"/avatars/d1747b2372e94500ecb5fb56809b482d.svg","isPro":false,"fullname":"Jinyeong Kim","user":"rubatoyeong","type":"user"},{"_id":"607f666a4ad99100d63ce35c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/607f666a4ad99100d63ce35c/QxhxnvfeV6efkxwUFHwjI.png","isPro":false,"fullname":"Bill Yuchen Lin","user":"yuchenlin","type":"user"},{"_id":"6486dde1f74857df3f1a5828","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6486dde1f74857df3f1a5828/FgE80CpalBO5qqArdfwxA.jpeg","isPro":false,"fullname":"Shenzhi Wang","user":"shenzhi-wang","type":"user"},{"_id":"655ac762cb17ec19ef82719b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/655ac762cb17ec19ef82719b/1kDncYrGLYS_2SR8cNdAL.png","isPro":false,"fullname":"Welcome to matlok","user":"matlok","type":"user"},{"_id":"649a46aea0b21c7cef780924","avatarUrl":"/avatars/47341742aaa7b8e4a39e67ceb81b28b0.svg","isPro":false,"fullname":"ddd","user":"ddxxasdf","type":"user"},{"_id":"644e1b1d9b4e87c31bab0a14","avatarUrl":"/avatars/88bb4c4a67dc8958069e9014f5e73a0b.svg","isPro":false,"fullname":"Michael Barry","user":"MichaelBarryUK","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"60078446e55258e41786a959","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/60078446e55258e41786a959/UGPCE4YqG9BVMSf0YauxL.png","isPro":false,"fullname":"Motoki Wu","user":"tokestermw","type":"user"},{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},{"_id":"62a4ac6fd83c3facafa50892","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62a4ac6fd83c3facafa50892/qFpobw9B5XaLZvwn0XbmB.jpeg","isPro":false,"fullname":"Mohammed Brıman","user":"mohammedbriman","type":"user"},{"_id":"656e3808d4de03a07d116850","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/656e3808d4de03a07d116850/62cFw46AmuhdI3gS24F1M.jpeg","isPro":true,"fullname":"Kenneth Hamilton","user":"ZennyKenny","type":"user"},{"_id":"6270324ebecab9e2dcf245de","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6270324ebecab9e2dcf245de/cMbtWSasyNlYc9hvsEEzt.jpeg","isPro":false,"fullname":"Kye Gomez","user":"kye","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2407.10457

The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore Non-Determinism

Published on Jul 15, 2024

· Submitted by

Bill Yuchen Lin on Jul 16, 2024

Upvote

Authors:

Yifan Song ,

Guoyin Wang ,

Bill Yuchen Lin

Abstract

Research explores non-determinism in LLM evaluations, comparing greedy decoding and sampling, and finds that smaller models can match larger ones with best-of-N sampling.

AI-generated summary

Current evaluations of large language models (LLMs) often overlook non-determinism, typically focusing on a single output per example. This limits our understanding of LLM performance variability in real-world applications. Our study addresses this issue by exploring key questions about the performance differences between greedy decoding and sampling, identifying benchmarks' consistency regarding non-determinism, and examining unique model behaviors. Through extensive experiments, we observe that greedy decoding generally outperforms sampling methods for most evaluated tasks. We also observe consistent performance across different LLM sizes and alignment methods, noting that alignment can reduce sampling variance. Moreover, our best-of-N sampling approach demonstrates that smaller LLMs can match or surpass larger models such as GPT-4-Turbo, highlighting the untapped potential of smaller LLMs. This research shows the importance of considering non-determinism in LLM evaluations and provides insights for future LLM development and evaluation.