LLMs still have a confidence calibration problem
danshipper · x · 2026-07-20
The post argues that LLMs often have a serious self-confidence problem: when asked how likely a life-changing event is, they may answer without giving a realistic sense of uncertainty.
The author points to a specific example presented as “a story in two photos,” using it to suggest that models still struggle with calibrated confidence rather than just raw capability.
More from Models
- Moonshot’s Kimi K3 reaches #5 on MathArena as the top open model — xeophon · 2026-07-22
- Google launches Gemini 3.5 Flash Cyber for CodeMender, with limited access for governments — GoogleAI · 2026-07-22
- Kimi K3 tops Gemini 3.6 Flash on four shared public benchmarks — ChrisGPT · 2026-07-22
- Google’s year-long pause in new base-model pretraining draws sharp criticism — teortaxesTex · 2026-07-22
- Current setup is 8,192 input tokens and 2,048 output tokens, with 8k/512 next — TheZachMueller · 2026-07-22
- Kimi K3 feels slower than K2.7, but stronger on long coding jobs and refactoring — Far-Presence2711 · 2026-07-22