5 papers
ReLE: A Scalable System and Structured Benchmark for Diagnosing Capability Anisotropy in Chinese LLMs
Rui Fang, Jian Li, Wei Chen +4
Large Language Models (LLMs) have achieved rapid progress in Chinese language understanding, yet accurately evaluating their capabilities remains challenged by benchmark saturation…
SSFO: Self-Supervised Faithfulness Optimization for Retrieval-Augmented Generation
Xiaqiang Tang, Yi Wang, Keyu Hu +5
Retrieval-Augmented Generation (RAG) systems require Large Language Models (LLMs) to generate responses that are faithful to the retrieved context. However, faithfulness hallucinat…
CogniBench: A Legal-inspired Framework and Dataset for Assessing Cognitive Faithfulness of Large Language Models
Xiaqiang Tang, Jian Li, Keyu Hu +5
Faithfulness hallucinations are claims generated by a Large Language Model (LLM) not supported by contexts provided to the LLM. Lacking assessment standards, existing benchmarks fo…
MBA-RAG: a Bandit Approach for Adaptive Retrieval-Augmented Generation through Question Complexity
Xiaqiang Tang, Qiang Gao, Jian Li +3
Retrieval Augmented Generation (RAG) has proven to be highly effective in boosting the generative performance of language model in knowledge-intensive tasks. However, existing RAG…
Adapting to Non-Stationary Environments: Multi-Armed Bandit Enhanced Retrieval-Augmented Generation on Knowledge Graphs
Xiaqiang Tang, Jian Li, Nan Du +1
Despite the superior performance of Large language models on many NLP tasks, they still face significant limitations in memorizing extensive world knowledge. Recent studies have de…