EMNLP Findings2023
Search Augmented Instruction Learning
Hongyin Luo, Tianhua Zhang, Yung-Sung Chuang, Yuan Gong, Yoon Kim, Xixin Wu, Helen Meng, James R. Glass
摘要
It is widely believed that connecting large language models with search engines can improve their transparency, truthfulness, and accessing to up-to-date information. However, we show that search grounding introduces new challenges to language models because of distracting, misleading, and untrustworthy information. To deal with these difficulties, we propose search-augmented instruction learning (SAIL), which allows a fine-tuned language model to source, denoise, and reason based on a mixed set of helpful and distracting search results. With an instruction tuning corpus, we collect search results for each training case from different search APIs and domains, and construct a new search-grounded training set containing (instruction, grounding information, response) triplets. We then fine-tune the LLaMA-7B model on the constructed training set. Since the collected search results contain distracting and disputing languages, the model needs to learn to ground on trustworthy search results, filter out distracting passages, and generate the target response. The search result-denoising process entails explicit trustworthy information selection and multi-hop reasoning, since the retrieved passages might be informative but not contain the instruction-following answer. Experiments show that the fine-tuned SAIL-7B model has a strong instruction-following ability, and it performs significantly better on transparencysensitive tasks, including question answering and fact checking.
问 Lune
几分钟读透这篇 Paper,不用再花一下午。
Lune 能读完全文,也能读引用它的顶会 Paper,讲清它的结论、证据和局限,每一点都注明出自哪篇 Paper。不用硬啃,也不靠猜。
已有 4 篇顶会 Paper 引用了它。别等审稿人问起,先弄清它们发现了什么。
脉络
- MMSearch-R1: Incentivizing LMMs to Search 引用了 Toolformer: Language Models Can Teach Themselves to Use Tools
- MMSearch-R1: Incentivizing LMMs to Search 引用了 RankRAG: Unifying Context Ranking with Retrieval-Augmented Generation in LLMs
- MMSearch-R1: Incentivizing LMMs to Search 引用了 Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models 引用了 Language Models are Few-Shot Learners
- PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models 引用了 Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- Language Models are Few-Shot Learners 引用了 Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- Shanks: Simultaneous Hearing and Thinking for Spoken Language Models 引用了 Toolformer: Language Models Can Teach Themselves to Use Tools
- Shanks: Simultaneous Hearing and Thinking for Spoken Language Models 引用了 Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
- Chain-of-Thought Prompting Elicits Reasoning in Large Language Models 引用了 Training language models to follow instructions with human feedback
- Training language models to follow instructions with human feedback 引用了 Language Models are Few-Shot Learners
- RankRAG: Unifying Context Ranking with Retrieval-Augmented Generation in LLMs 引用了 Training language models to follow instructions with human feedback
- RankRAG: Unifying Context Ranking with Retrieval-Augmented Generation in LLMs 引用了 Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
引用它的顶会 Paper4
- RankRAG: Unifying Context Ranking with Retrieval-Augmented Generation in LLMsYue Yu, Wei Ping, Zihan Liu, Boxin Wang 等NeurIPS 2024 · 321 次引用
- MMSearch-R1: Incentivizing LMMs to SearchJinming Wu, Zihao Deng, Wei Li, Yiding Liu 等ACL 2026 · 93 次引用
- Shanks: Simultaneous Hearing and Thinking for Spoken Language ModelsCheng-Han Chiang, Xiaofei Wang, Linjie Li, Chung-Ching Lin 等ACL 2026 · 14 次引用
- PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language ModelsWei Zou, Runpeng Geng, Binghui Wang, Jinyuan JiaUSENIX Security 2025
它引用的 Paper16
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah 等NeurIPS 2020 · 64,255 次引用
- Training language models to follow instructions with human feedbackLong Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida 等NeurIPS 2022 · 24,707 次引用
- Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma 等NeurIPS 2022 · 22,562 次引用
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni 等NeurIPS 2020 · 19,162 次引用
- Toolformer: Language Models Can Teach Themselves to Use ToolsTimo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu 等NeurIPS 2023 · 5,989 次引用
相关 Paper
- Fine-Tuning Language Models for FactualityKatherine Tian, Eric Mitchell, Huaxiu Yao, Christopher D. Manning 等ICLR 2024 · 270 次引用
- Search Arena: Analyzing Search-Augmented LLMsMihran Miroyan, Tsung-Han Wu, Logan King, Tianle Li 等ICLR 2026 · 32 次引用
- Beneficial Reasoning Behaviors in Agentic Search and Effective Training Methods to Obtain ThemJiahe Jin, Abhijay Sai Paladugu, Chenyan XiongACL Findings 2026
- Exploring the Impact of Instruction-Tuning on LLM's Susceptibility to MisinformationKyubeen Han, Junseo Jang, Hongjin Kim, Geunyeong Jeong 等ACL 2025
- Trustworthy Medical Question Answering: An Evaluation-Centric SurveyYinuo Wang, Baiyang Wang, Robert E. Mercer, Frank Rudzicz 等EMNLP 2025 · 2 次引用
