공대생 도전 일지

LLM 평가 코드 분석하기 - lm evaluation harness

https://github.com/EleutherAI/lm-evaluation-harness GitHub - EleutherAI/lm-evaluation-harness: A framework for few-shot evaluation of language models.A framework for few-shot evaluation of language models. - EleutherAI/lm-evaluation-harnessgithub.com https://deepwiki.com/EleutherAI/lm-evaluation-harness EleutherAI/lm-evaluation-harness | DeepWikiThe Language Model Evaluation Harness (LM-Eval) is..

L2M - Least-to-Most Prompting Enables Complex Reasoning in Large Language Models

https://arxiv.org/abs/2205.10625 Least-to-Most Prompting Enables Complex Reasoning in Large Language ModelsChain-of-thought prompting has demonstrated remarkable performance on various natural language reasoning tasks. However, it tends to perform poorly on tasks which requires solving problems harder than the exemplars shown in the prompts. To overcome this charxiv.org 앞에 작성했던 Decomposition 방식과..

Decomposed Prompting: A Modular Approach for Solving Complex Tasks

https://arxiv.org/abs/2210.02406 Decomposed Prompting: A Modular Approach for Solving Complex TasksFew-shot prompting is a surprisingly powerful way to use Large Language Models (LLMs) to solve various tasks. However, this approach struggles as the task complexity increases or when the individual reasoning steps of the task themselves are hard to learn,arxiv.org Decomposition의 시초격 논문인 것 같습니다. ..

To Believe or Not to Believe Your LLM

https://arxiv.org/abs/2406.02543 To Believe or Not to Believe Your LLMWe explore uncertainty quantification in large language models (LLMs), with the goal to identify when uncertainty in responses given a query is large. We simultaneously consider both epistemic and aleatoric uncertainties, where the former comes from the laarxiv.orgAU - 대답의 모호성EU - 학습의 부족 ! 이 논문도 결국 샘플링이라 해야 하나, 반복적인 모델 출력을 통해 ..

2025.05.03 - [인공지능/논문 리뷰 or 진행] - Planning 논문 모아 보기 NAACL 2025 - 3 Planning 논문 모아 보기 NAACL 2025 - 32025.05.02 - [인공지능/논문 리뷰 or 진행] - Agent, Hallucination 관련, Planning 논문 모아 보기 NAACL 2025 - 2 Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 22025.05.01 - [인공지능/논문 리뷰 or 진행] - Ayoonschallenge.tistory.com https://arxiv.org/abs/2503.17990 SUNAR: Semantic Uncertainty based Neighborhood Aware Re..

2025.05.02 - [인공지능/논문 리뷰 or 진행] - Agent, Hallucination 관련, Planning 논문 모아 보기 NAACL 2025 - 2 Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 22025.05.01 - [인공지능/논문 리뷰 or 진행] - Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 1 Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 1https://2025.naacl.org/program/accepted_papers/#main-coyoonschallenge.tistory.com이번엔 이어서 planning 관련..

2025.05.01 - [인공지능/논문 리뷰 or 진행] - Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 1 Agent, Hallucination 관련, Planning 논문 모아 보기 NACCL 2025 - 1https://2025.naacl.org/program/accepted_papers/#main-conference---long-papers Accepted PapersNAACL 2025 Accepted Papers2025.naacl.org진행되는 연구에서 논문을 찾아봐야 해서... https://arxiv.org/abs/2406.04784 SelfGoal: Your Language Agentsyoonschallenge.tistory.com이 ..

https://2025.naacl.org/program/accepted_papers/#main-conference---long-papers Accepted PapersNAACL 2025 Accepted Papers2025.naacl.org진행되는 연구에서 논문을 찾아봐야 해서... https://arxiv.org/abs/2406.04784 SelfGoal: Your Language Agents Already Know How to Achieve High-level GoalsLanguage agents powered by large language models (LLMs) are increasingly valuable as decision-making tools in domains such as gaming..

https://arxiv.org/abs/2408.00754 Coarse Correspondences Boost Spatial-Temporal Reasoning in Multimodal Language ModelMultimodal language models (MLLMs) are increasingly being applied in real-world environments, necessitating their ability to interpret 3D spaces and comprehend temporal dynamics. Current methods often rely on specialized architectural designs or task-speciarxiv.org기존 Multi-Modal은 ..

https://arxiv.org/abs/2503.15272 MAMM-Refine: A Recipe for Improving Faithfulness in Generation with Multi-Agent CollaborationMulti-agent collaboration among models has shown promise in reasoning tasks but is underexplored in long-form generation tasks like summarization and question-answering. We extend multi-agent multi-model reasoning to generation, specifically to improving farxiv.org 일단 여기서..