DiaHalu: A Dialogue-level Hallucination Evaluation Benchmark for Large Language Models
This addresses the need for more realistic and comprehensive hallucination evaluation in LLMs, though it is incremental as it builds on existing benchmarks by extending to dialogue-level.
The authors tackled the problem of evaluating hallucinations in large language models by introducing DiaHalu, the first dialogue-level benchmark covering multiple domains and hallucination subtypes, with experiments showing it is a challenging dataset for further research.
Since large language models (LLMs) achieve significant success in recent years, the hallucination issue remains a challenge, numerous benchmarks are proposed to detect the hallucination. Nevertheless, some of these benchmarks are not naturally generated by LLMs but are intentionally induced. Also, many merely focus on the factuality hallucination while ignoring the faithfulness hallucination. Additionally, although dialogue pattern is more widely utilized in the era of LLMs, current benchmarks only concentrate on sentence-level and passage-level hallucination. In this study, we propose DiaHalu, the first dialogue-level hallucination evaluation benchmark to our knowledge. Initially, we integrate the collected topics into system prompts and facilitate a dialogue between two ChatGPT3.5. Subsequently, we manually modify the contents that do not adhere to human language conventions and then have LLMs re-generate, simulating authentic human-machine interaction scenarios. Finally, professional scholars annotate all the samples in the dataset. DiaHalu covers four common multi-turn dialogue domains and five hallucination subtypes, extended from factuality and faithfulness hallucination. Experiments through some well-known LLMs and detection methods on the dataset show that DiaHalu is a challenging benchmark, holding significant value for further research.