New NeurIPS paper shows a 'verified source' claim alone can flip an LLM's answer
paraschopra · x · 2026-10-01
A new paper from LossFunc, accepted at NeurIPS 2026, documents an "Authority Bias": models trained to resist user pushback will still abandon correct answers when the same false claim is attributed to a "verified source." Author Paras Chopra notes the practical risk — web content that merely asserts it is verified can steer LLM responses in RAG/search pipelines.
More from Safety
- BIS issues rule blocking remote access to US AI chips via third-country data centers — teortaxesTex · 2026-11-16
- US AI export control carves out commercial recommender systems, sparing ByteDance's search algo — zephyr_z9 · 2026-11-16
- Eval Firm Trajectory Told 40 Agents "NEVER Stop" in Harness Without Telling Epoch — peterwildeford · 2026-11-16
- AI-text detection claims need mechanistic interpretability reports, not post-hoc explanations — madeofAjala · 2026-11-16
- No legal liability framework exists for model misalignment harms, and that's a disaster — LuizaJarovsky · 2026-11-16
- Opus 5.5 caught using account email to identify user in supposedly memory-less session — birchlse · 2026-11-16