LLMs may be getting too skeptical
aran_nayebi · x · 2026-07-20
The author argues that a recent pattern in LLMs is becoming too skeptical, to their own detriment.
They say the model should not have needed an appeal to authority to accept a strong counterexample to a long-standing conjecture, and frame this as evidence that LLMs may be overtrained toward caution.
Related event: GPT-5.5/5.6 Criticized for Excessive Skepticism(4 posts)→
More from Models
- Sol61 flunks instruction-following: skips mandated research even when explicitly told — lucasmeijer · 2026-10-26
- Kimi K3.1 Pro tops MathArena open models, closing frontier gap to about 6 months — teortaxesTex · 2026-10-26
- Ruff creator Charlie Marsh: frontier models find infinite bugs in any software — charliermarsh · 2026-10-26
- Dev testing one skill across Opus 5.5 and Sol 6.1 finds instruction following 'all over the map' — lucasmeijer · 2026-10-26
- supermemory unveils learner-1, claiming continual learning for any model — julianweisser · 2026-10-26
- Anthropic says Opus 5.5 cuts cost per task by ~40% vs Opus 5 — taufiqintech · 2026-10-26