JPMorgan Chase publishes application on LLM hallucination detection and fine-tuning.
The filing
JPMorgan Chase published an application on November 12, 2026, describing systems and methods to improve large language model reliability through hallucination-averse fine-tuning and detection. The applicant describes the approach as enhancing reliability by detecting and correcting hallucinations in large language models. US20260359628A1 (opens in a new tab)
This filing’s full text wasn’t read for this issue, so its claims and the applicant’s background aren’t shown. Both are in the record, under Sources.