Deprecated: The each() function is deprecated. This message will be suppressed on further calls in /home/zhenxiangba/zhenxiangba.com/public_html/phproxy-improved-master/index.php on line 456 Paper page - Lemur: Harmonizing Natural Language and Code for Language Agents
Full summary is here. Paper is here.\n","updatedAt":"2023-10-13T14:20:16.537Z","author":{"_id":"6486638da4cf2081f20c40ec","avatarUrl":"/avatars/0bc16a7447cd71ac18828a678313bd83.svg","fullname":"Mike Young","name":"mikelabs","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":13,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8945189118385315},"editors":["mikelabs"],"editorAvatarUrls":["/avatars/0bc16a7447cd71ac18828a678313bd83.svg"],"reactions":[{"reaction":"๐ค","users":["mikelabs","victor","davanstrien","tianbaoxiexxx","osanseviero"],"count":5}],"isReport":false}},{"id":"652a6d4660e70673058e549c","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":"2023-10-14T10:28:22.000Z","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* [MINT: Evaluating LLMs in Multi-turn Interaction with Tools and Language Feedback](https://huggingface.co/papers/2309.10691) (2023)\n* [L2CEval: Evaluating Language-to-Code Generation Capabilities of Large Language Models](https://huggingface.co/papers/2309.17446) (2023)\n* [Qwen Technical Report](https://huggingface.co/papers/2309.16609) (2023)\n* [At Which Training Stage Does Code Data Help LLMs Reasoning?](https://huggingface.co/papers/2309.16298) (2023)\n* [FireAct: Toward Language Agent Fine-tuning](https://huggingface.co/papers/2310.05915) (2023)\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","html":"
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
\n
The following papers were recommended by the Semantic Scholar API
\n","updatedAt":"2024-06-09T07:19:46.909Z","author":{"_id":"6186ddf6a7717cb375090c01","avatarUrl":"/avatars/716b6a7d1094c8036b2a8a7b9063e8aa.svg","fullname":"Julien BLANCHON","name":"blanchon","type":"user","isPro":true,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":176,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5176085233688354},"editors":["blanchon"],"editorAvatarUrls":["/avatars/716b6a7d1094c8036b2a8a7b9063e8aa.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2310.06830","authors":[{"_id":"652864bf1ddb701dc2cba3ab","user":{"_id":"601d29ab913ad3afd7b7ddb8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1620447944896-601d29ab913ad3afd7b7ddb8.jpeg","isPro":false,"fullname":"Yiheng Xu","user":"ranpox","type":"user"},"name":"Yiheng Xu","status":"claimed_verified","statusLastChangedAt":"2023-10-13T07:52:25.749Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3ac","user":{"_id":"632d53951538d4798a73c849","avatarUrl":"/avatars/b7d0a895e669bcd1303c4716b5401c36.svg","isPro":false,"fullname":"Hongjin SU","user":"multi-train","type":"user"},"name":"Hongjin Su","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:33:10.404Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3ad","user":{"_id":"649e2bfbab373ff26ac20853","avatarUrl":"/avatars/0df2ada953e4553b773ff064dd86228c.svg","isPro":false,"fullname":"Chen Xing","user":"LynetteXing","type":"user"},"name":"Chen Xing","status":"claimed_verified","statusLastChangedAt":"2023-10-13T14:43:01.179Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3ae","user":{"_id":"64858110b28e374571bbc550","avatarUrl":"/avatars/62c169a3149d5a7942c960753df3a673.svg","isPro":false,"fullname":"Mi Boyu","user":"miboyu5","type":"user"},"name":"Boyu Mi","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:33:50.091Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3af","user":{"_id":"612ee6a7b960e78c6d2319d4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/612ee6a7b960e78c6d2319d4/2Hu9BaAyXbyh1vt0v1Qui.jpeg","isPro":false,"fullname":"Qian Liu","user":"SivilTaram","type":"user"},"name":"Qian Liu","status":"claimed_verified","statusLastChangedAt":"2023-10-13T07:52:22.016Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b0","user":{"_id":"6400f2ed568dbe30c9161e47","avatarUrl":"/avatars/c55938df5bce82b5d96e592a1ec36a8b.svg","isPro":false,"fullname":"Weijia Shi","user":"swj0419","type":"user"},"name":"Weijia Shi","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:34:18.056Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b1","user":{"_id":"61e4c4ca1ab24785ac11ba69","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/61e4c4ca1ab24785ac11ba69/1Q1zhhyGSJ9RJG9MzwxVv.jpeg","isPro":false,"fullname":"Binyuan Hui","user":"huybery","type":"user"},"name":"Binyuan Hui","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:34:25.742Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b2","name":"Fan Zhou","hidden":false},{"_id":"652864bf1ddb701dc2cba3b3","user":{"_id":"6482a35b93362a0d12172f9a","avatarUrl":"/avatars/8690151b24da63521de6abf7dd65838d.svg","isPro":false,"fullname":"Yitao Liu","user":"YitaoLiu","type":"user"},"name":"Yitao Liu","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:35:11.354Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b4","user":{"_id":"618767e4238063b4615d042b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1636263880877-noauth.jpeg","isPro":false,"fullname":"Tianbao Xie","user":"tianbaoxiexxx","type":"user"},"name":"Tianbao Xie","status":"claimed_verified","statusLastChangedAt":"2023-10-13T07:52:18.792Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b5","user":{"_id":"6083902e1e36b13a64497d91","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6083902e1e36b13a64497d91/h4rGHMn2c6z5GesF0F6VU.png","isPro":false,"fullname":"cheng","user":"zhoujun","type":"user"},"name":"Zhoujun Cheng","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:35:24.367Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b6","user":{"_id":"642e76a8113b200fd767b3c9","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/642e76a8113b200fd767b3c9/03_CAAr6gy7H_F_iUWx76.png","isPro":false,"fullname":"Siheng Zhao","user":"OliverZhao","type":"user"},"name":"Siheng Zhao","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:35:32.212Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b7","name":"Lingpeng Kong","hidden":false},{"_id":"652864bf1ddb701dc2cba3b8","user":{"_id":"642f4bfbc953ca48ccd162f2","avatarUrl":"/avatars/390662485bbf2bbe3454ba9f416d1536.svg","isPro":false,"fullname":"Bailin","user":"bailin28","type":"user"},"name":"Bailin Wang","status":"claimed_verified","statusLastChangedAt":"2023-10-16T07:24:53.075Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3b9","user":{"_id":"649dbcc4e0fff1ed099dc80a","avatarUrl":"/avatars/c87c273ca628dbcddccbf1ee19b2ce33.svg","isPro":false,"fullname":"Caiming Xiong","user":"cxiong","type":"user"},"name":"Caiming Xiong","status":"admin_assigned","statusLastChangedAt":"2023-10-13T08:36:36.549Z","hidden":false},{"_id":"652864bf1ddb701dc2cba3ba","name":"Tao Yu","hidden":false}],"publishedAt":"2023-10-10T17:57:45.000Z","submittedOnDailyAt":"2023-10-13T04:13:33.787Z","title":"Lemur: Harmonizing Natural Language and Code for Language Agents","submittedOnDailyBy":{"_id":"60f1abe7544c2adfd699860c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674929746905-60f1abe7544c2adfd699860c.jpeg","isPro":false,"fullname":"AK","user":"akhaliq","type":"user"},"summary":"We introduce Lemur and Lemur-Chat, openly accessible language models\noptimized for both natural language and coding capabilities to serve as the\nbackbone of versatile language agents. The evolution from language chat models\nto functional language agents demands that models not only master human\ninteraction, reasoning, and planning but also ensure grounding in the relevant\nenvironments. This calls for a harmonious blend of language and coding\ncapabilities in the models. Lemur and Lemur-Chat are proposed to address this\nnecessity, demonstrating balanced proficiencies in both domains, unlike\nexisting open-source models that tend to specialize in either. Through\nmeticulous pre-training using a code-intensive corpus and instruction\nfine-tuning on text and code data, our models achieve state-of-the-art averaged\nperformance across diverse text and coding benchmarks among open-source models.\nComprehensive experiments demonstrate Lemur's superiority over existing\nopen-source models and its proficiency across various agent tasks involving\nhuman communication, tool usage, and interaction under fully- and partially-\nobservable environments. The harmonization between natural and programming\nlanguages enables Lemur-Chat to significantly narrow the gap with proprietary\nmodels on agent abilities, providing key insights into developing advanced\nopen-source agents adept at reasoning, planning, and operating seamlessly\nacross environments. https://github.com/OpenLemur/Lemur","upvotes":33,"discussionId":"652864c01ddb701dc2cba3fd","githubRepo":"https://github.com/openlemur/lemur","githubRepoAddedBy":"auto","ai_summary":"Lemur and Lemur-Chat are open-source language models combining language and coding capabilities, achieving state-of-the-art performance across various benchmarks and agent tasks.","ai_keywords":["language models","coding capabilities","harmonious blend","language and coding","pre-training","code-intensive corpus","instruction fine-tuning","state-of-the-art","diverse text and coding benchmarks","agent tasks","human communication","tool usage","interaction","fully-observable environments","partially-observable environments","proprietary models","advanced open-source agents"],"githubStars":555},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"612ee6a7b960e78c6d2319d4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/612ee6a7b960e78c6d2319d4/2Hu9BaAyXbyh1vt0v1Qui.jpeg","isPro":false,"fullname":"Qian Liu","user":"SivilTaram","type":"user"},{"_id":"601d29ab913ad3afd7b7ddb8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1620447944896-601d29ab913ad3afd7b7ddb8.jpeg","isPro":false,"fullname":"Yiheng Xu","user":"ranpox","type":"user"},{"_id":"618767e4238063b4615d042b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1636263880877-noauth.jpeg","isPro":false,"fullname":"Tianbao Xie","user":"tianbaoxiexxx","type":"user"},{"_id":"63c9473b00104ea998dd5b6b","avatarUrl":"/avatars/3f73ed8e73a99d61722255793718c018.svg","isPro":false,"fullname":"Luoxuan Weng","user":"WhiteWolf82","type":"user"},{"_id":"6083902e1e36b13a64497d91","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6083902e1e36b13a64497d91/h4rGHMn2c6z5GesF0F6VU.png","isPro":false,"fullname":"cheng","user":"zhoujun","type":"user"},{"_id":"61e4c4ca1ab24785ac11ba69","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/61e4c4ca1ab24785ac11ba69/1Q1zhhyGSJ9RJG9MzwxVv.jpeg","isPro":false,"fullname":"Binyuan Hui","user":"huybery","type":"user"},{"_id":"611a31e94e5e661f8b21ebbf","avatarUrl":"/avatars/d7e4ce3cbe71bc7cbbab454a087c265c.svg","isPro":false,"fullname":"Tristan Marechaux","user":"tmarechaux","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":"642e76a8113b200fd767b3c9","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/642e76a8113b200fd767b3c9/03_CAAr6gy7H_F_iUWx76.png","isPro":false,"fullname":"Siheng Zhao","user":"OliverZhao","type":"user"},{"_id":"5f17f0a0925b9863e28ad517","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/5f17f0a0925b9863e28ad517/fXIY5i9RLsIa1v3CCuVtt.jpeg","isPro":true,"fullname":"Victor Mustar","user":"victor","type":"user"},{"_id":"5dd96eb166059660ed1ee413","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/5dd96eb166059660ed1ee413/NQtzmrDdbG0H8qkZvRyGk.jpeg","isPro":true,"fullname":"Julien Chaumond","user":"julien-c","type":"user"},{"_id":"6090856b1e62cfa4f5c23ccb","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6090856b1e62cfa4f5c23ccb/XtetET8dL65viJDOKIWzl.jpeg","isPro":true,"fullname":"Thomas Wood","user":"odellus","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":2}">
Lemur and Lemur-Chat are open-source language models combining language and coding capabilities, achieving state-of-the-art performance across various benchmarks and agent tasks.
AI-generated summary
We introduce Lemur and Lemur-Chat, openly accessible language models
optimized for both natural language and coding capabilities to serve as the
backbone of versatile language agents. The evolution from language chat models
to functional language agents demands that models not only master human
interaction, reasoning, and planning but also ensure grounding in the relevant
environments. This calls for a harmonious blend of language and coding
capabilities in the models. Lemur and Lemur-Chat are proposed to address this
necessity, demonstrating balanced proficiencies in both domains, unlike
existing open-source models that tend to specialize in either. Through
meticulous pre-training using a code-intensive corpus and instruction
fine-tuning on text and code data, our models achieve state-of-the-art averaged
performance across diverse text and coding benchmarks among open-source models.
Comprehensive experiments demonstrate Lemur's superiority over existing
open-source models and its proficiency across various agent tasks involving
human communication, tool usage, and interaction under fully- and partially-
observable environments. The harmonization between natural and programming
languages enables Lemur-Chat to significantly narrow the gap with proprietary
models on agent abilities, providing key insights into developing advanced
open-source agents adept at reasoning, planning, and operating seamlessly
across environments. https://github.com/OpenLemur/Lemur
Yay, papers are back! Here's my summary of this article.
Today's conversational bots like Claude and GPT can chat impressively but aren't great at complex planning or executing technical tasks. To overcome this, new research from HKU builds open-source AI agents that blend natural language and coding skills. They're called Lemur and Lemur-Chat.
The researchers think achieving versatile real-world agents requires models that integrate both fluid natural language abilities and precise programming language control. Humans combine plain speech for higher-level goals with languages like Python when we need to plan intricately and execute exactly. AI needs both capacities too.
But most existing models specialize in pure language or pure code. There's a separation that is limiting.
The team created Lemur by pretraining the open-source Llama-2 on a massive mixed corpus with 10x more natural language than code. This improved its programming abilities while retaining conversational strength. Further instruction tuning optimized Lemur-Chat for following free-form directions in language.
Experiments found Lemur surpassed specialized coding-only models like Codex in overall benchmarks. Lemur-Chat then exceeded Lemur by 15% after instruction tuning.
More importantly, Lemur-Chat won 12/13 new "agent tests" designed to mimic real-world challenges needing both language and programming prowess.
It beat alternatives at:
Using tools like Python and Wikipedia to enhance reasoning
Debugging code by leveraging error messages
Improving the most from natural language feedback
Exploring partially observable environments like cybersecurity and web browsing simulations.
Lemur-Chat matched GPT-3.5 in many tests, closing the gap between commercial and open-source agents.
TLDR: New open-source AI agents combine coding and language skills. Experiments show the combo unlocks more performance across technical challenges.