AI Paper Atlas

Research map / Large language models / Alignment and preferences

LIMA: Less Is More for Alignment

Chunting Zhou, Pengfei Liu, Puxin Xu, Srini Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, LILI YU, Susan Zhang, Gargi Ghosh and 3 more

NeurIPS 2023 · Large language models › Alignment and preferences · approach: Steering and correcting LLM behavior · 128 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use diverse prompts rather than homogeneous how-to prompts (ablation)LLaMa 7B fine-tuned on 2,000 examples · Stack Exchange versus wikiHowChatGPT helpfulness scoreImproved
Use quality-filtered responses rather than unfiltered responses (ablation)LLaMa 7B fine-tuned on 2,000 Stack Exchange examples · Stack ExchangeChatGPT helpfulness score: 0.5 point differenceImproved
Increase training-set quantity without increasing prompt diversity (ablation)LLaMa 7B trained on quality-filtered Stack Exchange · Stack ExchangeChatGPT helpfulness score: up to a 16-fold increaseNo clear effect
Add 30 multi-turn dialogue chains (ablation)LIMA trained from pretrained LLaMa · 10 live conversationsexcellent responses: 45.2% to 76.1%Improved
Add 30 multi-turn dialogue chains (ablation)LIMA trained from pretrained LLaMa · 10 live conversationsfailed turns: 15 fails per 42 turns to 1 fail per 46Improved
Add six examples with varied formatting constraints (ablation)LIMA · Prompts requiring complex structureresponse qualityImproved
Fine-tune on 1,000 curated examples rather than using only the pretrained modelLLaMa 65B · 300 test promptshuman preference versus GPT-4: 43%Mixed
Fine-tune on 1,000 curated examples rather than using only the pretrained modelLLaMa 65B · 300 test promptshuman preference versus DaVinci003: 65%Improved
Train on 1,000 curated examples rather than 52,000 Alpaca examplesLLaMa 65B · 300 test promptshuman preferenceImproved

Later papers that compared with this method

4 later accepted papers report LIMA as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy