NYU Researchers Challenge Anthropic's Claim That LLMs Can Introspect
tallinzen · x · 2026-10-03
Anthropic previously published work arguing that LLMs can introspect—recognizing and reporting on their own internal states. New research by NYU CDS PhD student Shashwat Singh, Associate Professor Tal Linzen, and Faculty Fellow Shauli Ravfogel finds the evidence falls short of supporting that claim. The paper is being presented at a poster session, with an accessible writeup on the NYU CDS blog.
Related event: Researchers to Challenge Anthropic's Claims of LLM Introspection at COLM(3 posts)→
More from Models
- RSIArena, 48 Hours In: Model Catches Test Questions in Its Training Data and Restarts — my_cat_can_code · 2026-10-03
- Spotlight architecture scrutiny: attention-style memory may carry a large constant factor — teortaxesTex · 2026-10-03
- GPT-6.1 Sol reportedly under heavy load; capacity expansion to nearly double serving speed — NandaVegg · 2026-10-03
- One prompt with Opus 5.5 builds a realistic fighter jet game — TAbrodi · 2026-10-03
- Perplexity drops seven open-source projects: SoTA decision model, on-device PII filter, security tools — AravSrinivas · 2026-10-03
- Are hallucinations improving or plateauing? Reddit thread probes AI reliability ceiling — maedhros256 · 2026-10-03