RECALL: A Benchmark for LLMs Robustness against External Counterfactual Knowledge

2023/11/14 by Yi Liu, Liu, Yi, Lianzhe Huang +13 · 11 citations
Computer Science · Medicine · #AI in Service Interactions #Artificial Intelligence (cs.AI) #Artificial Intelligence in Healthcare and Education #Computation and Language (cs.CL) #FOS: Computer and information sciences #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2311.08147

openalex publication_date 2023/11/14 · openalex created_date 2023/11/16 · openalex updated_date 2026/07/28

Abstract

LLMs and AI chatbots have improved people's efficiency in various fields. However, the necessary knowledge for answering the question may be beyond the models' knowledge boundaries. To mitigate this issue, many researchers try to introduce external knowledge, such as knowledge graphs and Internet contents, into LLMs for up-to-date information. However, the external information from the Internet may include counterfactual information that will confuse the model and lead to an incorrect response. Thus there is a pressing need for LLMs to possess the ability to distinguish reliable information from external knowledge. Therefore, to evaluate the ability of LLMs to discern the reliability of external knowledge, we create a benchmark from existing knowledge bases. Our benchmark consists of two tasks, Question Answering and Text Generation, and for each task, we provide models with a context containing counterfactual information. Evaluation results show that existing LLMs are susceptible to interference from unreliable external knowledge with counterfactual information, and simple intervention methods make limited contributions to the alleviation of this issue.

Cited by

• RAGalyst: Automated Human-Aligned Agentic Evaluation for Domain-Specific RAG
• Balancing Knowledge Updates: Toward Unified Modular Editing in LLMs
• Memory in Large Language Models: Mechanisms, Evaluation and Evolution
• Can we Evaluate RAGs with Synthetic Data?
• "Lost-in-the-Later": Framework for Quantifying Contextual Grounding in Large Language Models
• GaRAGe: A Benchmark with Grounding Annotations for RAG Evaluation
• Magic Mushroom: A Customizable Benchmark for Fine-grained Analysis of Retrieval Noise Erosion in RAG Systems
• Benchmark Test-Time Scaling of General LLM Agents
• MIRAGE: A Metric-Intensive Benchmark for Retrieval-Augmented Generation Evaluation
• Retrieval Augmented Generation Evaluation in the Era of Large Language Models: A Comprehensive Survey
• QE-RAG: A Robust Retrieval-Augmented Generation Benchmark for Query Entry Errors

Related

• Are LLMs Really Not Knowledgeable? Mining the Submerged Knowledge in LLMs' Memory
• LLMs as Repositories of Factual Knowledge: Limitations and Solutions
• Rejection Improves Reliability: Training LLMs to Refuse Unknown Questions Using RL from Knowledge Feedback
• Know Or Not: a library for evaluating out-of-knowledge base robustness
• How Reliable are LLMs as Knowledge Bases? Re-thinking Facutality and Consistency