Transferable and Efficient Non-Factual Content Detection via Probe Training with Offline Consistency Checking
This addresses the issue of trustworthiness in LLM generations for users relying on accurate information, though it appears incremental by building on existing consistency-checking approaches.
The paper tackles the problem of detecting non-factual content in large language model outputs by proposing PINOSE, which trains a probing model using offline self-consistency checking to avoid human annotations and reduce computational costs, achieving superior results on factuality detection and question answering benchmarks compared to existing methods.
Detecting non-factual content is a longstanding goal to increase the trustworthiness of large language models (LLMs) generations. Current factuality probes, trained using humanannotated labels, exhibit limited transferability to out-of-distribution content, while online selfconsistency checking imposes extensive computation burden due to the necessity of generating multiple outputs. This paper proposes PINOSE, which trains a probing model on offline self-consistency checking results, thereby circumventing the need for human-annotated data and achieving transferability across diverse data distributions. As the consistency check process is offline, PINOSE reduces the computational burden of generating multiple responses by online consistency verification. Additionally, it examines various aspects of internal states prior to response decoding, contributing to more effective detection of factual inaccuracies. Experiment results on both factuality detection and question answering benchmarks show that PINOSE achieves surpassing results than existing factuality detection methods. Our code and datasets are publicly available on this anonymized repository.