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

\n","updatedAt":"2023-12-06T16:03:53.541Z","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":0.7504582405090332},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[{"reaction":"👍","users":["bkhmsi","STARBORN"],"count":2}],"isReport":false}},{"id":"65e2bba701d1420e34871dfa","author":{"_id":"621d8a8d411b802eb0b70615","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1646104902183-621d8a8d411b802eb0b70615.png","fullname":"paoladimaio","name":"STARBORN","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":6,"isUserFollowing":false},"createdAt":"2024-03-02T05:39:51.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"VERY HELPFUL :-)","html":"

VERY HELPFUL :-)

\n","updatedAt":"2024-03-02T05:39:51.015Z","author":{"_id":"621d8a8d411b802eb0b70615","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1646104902183-621d8a8d411b802eb0b70615.png","fullname":"paoladimaio","name":"STARBORN","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":6,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"nl","probability":0.2589787542819977},"editors":["STARBORN"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1646104902183-621d8a8d411b802eb0b70615.png"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2312.00575","authors":[{"_id":"656d33b60bbc114fe639bcc8","user":{"_id":"631ad4b2fac58c9c8167bafc","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/631ad4b2fac58c9c8167bafc/jum1yrhVWkl-biBdKitF3.png","isPro":false,"fullname":"Khai Loong Aw","user":"awwkl","type":"user"},"name":"Khai Loong Aw","status":"admin_assigned","statusLastChangedAt":"2023-12-04T10:45:25.821Z","hidden":false},{"_id":"656d33b60bbc114fe639bcc9","user":{"_id":"6447a6e94ca52cdb3188adec","avatarUrl":"/avatars/624084372445299bf8d3a8751da93475.svg","isPro":false,"fullname":"Montariol","user":"Syrielle","type":"user"},"name":"Syrielle Montariol","status":"admin_assigned","statusLastChangedAt":"2023-12-04T10:45:35.553Z","hidden":false},{"_id":"656d33b60bbc114fe639bcca","user":{"_id":"63ca94752ffd350b486354e1","avatarUrl":"/avatars/2e41b2651d30b855f0b46cddaa02aa43.svg","isPro":false,"fullname":"Badr AlKhamissi","user":"bkhmsi","type":"user"},"name":"Badr AlKhamissi","status":"admin_assigned","statusLastChangedAt":"2023-12-04T10:45:44.223Z","hidden":false},{"_id":"656d33b60bbc114fe639bccb","user":{"_id":"62c23d7b99c6c7738a66989b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1656896923660-62c23d7b99c6c7738a66989b.jpeg","isPro":false,"fullname":"Martin Schrimpf","user":"mschrimpf","type":"user"},"name":"Martin Schrimpf","status":"admin_assigned","statusLastChangedAt":"2023-12-04T10:45:53.586Z","hidden":false},{"_id":"656d33b60bbc114fe639bccc","user":{"_id":"654fb1c67490049d621516fb","avatarUrl":"/avatars/0f88177da236b1bce4a4a58fd0246884.svg","isPro":false,"fullname":"Antoine Bosselut","user":"atcbosselut","type":"user"},"name":"Antoine Bosselut","status":"admin_assigned","statusLastChangedAt":"2023-12-04T10:45:59.665Z","hidden":false}],"publishedAt":"2023-12-01T13:31:02.000Z","submittedOnDailyAt":"2023-12-03T23:34:39.122Z","title":"Instruction-tuning Aligns LLMs to the Human Brain","submittedOnDailyBy":{"_id":"60f1abe7544c2adfd699860c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674929746905-60f1abe7544c2adfd699860c.jpeg","isPro":false,"fullname":"AK","user":"akhaliq","type":"user"},"summary":"Instruction-tuning is a widely adopted method of finetuning that enables\nlarge language models (LLMs) to generate output that more closely resembles\nhuman responses to natural language queries, in many cases leading to\nhuman-level performance on diverse testbeds. However, it remains unclear\nwhether instruction-tuning truly makes LLMs more similar to how humans process\nlanguage. We investigate the effect of instruction-tuning on LLM-human\nsimilarity in two ways: (1) brain alignment, the similarity of LLM internal\nrepresentations to neural activity in the human language system, and (2)\nbehavioral alignment, the similarity of LLM and human behavior on a reading\ntask. We assess 25 vanilla and instruction-tuned LLMs across three datasets\ninvolving humans reading naturalistic stories and sentences. We discover that\ninstruction-tuning generally enhances brain alignment by an average of 6%, but\ndoes not have a similar effect on behavioral alignment. To identify the factors\nunderlying LLM-brain alignment, we compute correlations between the brain\nalignment of LLMs and various model properties, such as model size, various\nproblem-solving abilities, and performance on tasks requiring world knowledge\nspanning various domains. Notably, we find a strong positive correlation\nbetween brain alignment and model size (r = 0.95), as well as performance on\ntasks requiring world knowledge (r = 0.81). Our results demonstrate that\ninstruction-tuning LLMs improves both world knowledge representations and brain\nalignment, suggesting that mechanisms that encode world knowledge in LLMs also\nimprove representational alignment to the human brain.","upvotes":15,"discussionId":"656d33b70bbc114fe639bcf1","ai_summary":"Instruction-tuning improves brain alignment in large language models but not behavioral alignment, with correlations found between brain alignment and model size and world knowledge performance.","ai_keywords":["instruction-tuning","large language models","brain alignment","behavioral alignment","neural activity","world knowledge","model size"]},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"651d618a18be7acf8e602c41","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/kEDoJKsGXpNDTOiU7FRMP.jpeg","isPro":false,"fullname":"Abreu Magalhães","user":"Hildeberto","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":"5e6a3d4ea9afd5125d9ec064","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1584020801691-noauth.jpeg","isPro":true,"fullname":"Stefan Schweter","user":"stefan-it","type":"user"},{"_id":"6549135c196ae037a74e10a3","avatarUrl":"/avatars/86194456844c7b2b5389de36cb258472.svg","isPro":false,"fullname":"Richrich","user":"RichardForests","type":"user"},{"_id":"63e4038e789dcaae43c52fd8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e4038e789dcaae43c52fd8/ElFZW_r9n_D6kLRL9IIlh.jpeg","isPro":false,"fullname":"Hari Koduvely","user":"harik68","type":"user"},{"_id":"6427532b39c7e60c4b37cc4b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6427532b39c7e60c4b37cc4b/XIR5R3oB4PcPIJLN9oYHD.jpeg","isPro":false,"fullname":"dann 🏴‍☠️","user":"dannoncaffeine","type":"user"},{"_id":"6538119803519fddb4a17e10","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6538119803519fddb4a17e10/ffJMkdx-rM7VvLTCM6ri_.jpeg","isPro":false,"fullname":"samusenps","user":"samusenps","type":"user"},{"_id":"64ca7c04710645aa7bdbbfff","avatarUrl":"/avatars/c12f4cb6dc1ff0010edb3ef4cfcccd7c.svg","isPro":false,"fullname":"Lize Pirenne","user":"Inversta","type":"user"},{"_id":"6141a88b3a0ec78603c9e784","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6141a88b3a0ec78603c9e784/DJsxSmWV39M33JFheLobC.jpeg","isPro":true,"fullname":"merve","user":"merve","type":"user"},{"_id":"643efec9e9d063936911026c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/643efec9e9d063936911026c/25TPUXWzFyBtdr7iH-T25.jpeg","isPro":false,"fullname":"Promptmetheus","user":"azure-arc-0","type":"user"},{"_id":"631ad4b2fac58c9c8167bafc","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/631ad4b2fac58c9c8167bafc/jum1yrhVWkl-biBdKitF3.png","isPro":false,"fullname":"Khai Loong Aw","user":"awwkl","type":"user"},{"_id":"663ccbff3a74a20189d4aa2e","avatarUrl":"/avatars/83a54455e0157480f65c498cd9057cf2.svg","isPro":false,"fullname":"Nguyen Van Thanh","user":"NguyenVanThanhHust","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0}">

Papers

arxiv:2312.00575

Instruction-tuning Aligns LLMs to the Human Brain

Published on Dec 1, 2023

· Submitted by

AK on Dec 3, 2023

Upvote

Authors:

Khai Loong Aw ,

Syrielle Montariol ,

Badr AlKhamissi ,

Martin Schrimpf ,

Antoine Bosselut

Abstract

Instruction-tuning improves brain alignment in large language models but not behavioral alignment, with correlations found between brain alignment and model size and world knowledge performance.

AI-generated summary

Instruction-tuning is a widely adopted method of finetuning that enables large language models (LLMs) to generate output that more closely resembles human responses to natural language queries, in many cases leading to human-level performance on diverse testbeds. However, it remains unclear whether instruction-tuning truly makes LLMs more similar to how humans process language. We investigate the effect of instruction-tuning on LLM-human similarity in two ways: (1) brain alignment, the similarity of LLM internal representations to neural activity in the human language system, and (2) behavioral alignment, the similarity of LLM and human behavior on a reading task. We assess 25 vanilla and instruction-tuned LLMs across three datasets involving humans reading naturalistic stories and sentences. We discover that instruction-tuning generally enhances brain alignment by an average of 6%, but does not have a similar effect on behavioral alignment. To identify the factors underlying LLM-brain alignment, we compute correlations between the brain alignment of LLMs and various model properties, such as model size, various problem-solving abilities, and performance on tasks requiring world knowledge spanning various domains. Notably, we find a strong positive correlation between brain alignment and model size (r = 0.95), as well as performance on tasks requiring world knowledge (r = 0.81). Our results demonstrate that instruction-tuning LLMs improves both world knowledge representations and brain alignment, suggesting that mechanisms that encode world knowledge in LLMs also improve representational alignment to the human brain.