Community probe: original gpt-oss knows nothing of the 'hidden classifier' fine-tune
repligate · x · 2026-11-06
A forwarded discussion offers a tell: ask the original open-source gpt-oss model about certain policies and it has no idea what you mean, while a fine-tuned classifier variant replies promptly about the policies. Both are accessible on OpenRouter, making the claim independently checkable — an unverified but reproducible observation suggesting undisclosed fine-tuned variants may be in circulation.
More from Models
- METR-style fit shows benchmark gains accelerating 1.9x, but sensitive to measurement — jessi_cata · 2026-11-06
- 7 of Anthropic's last 8 frontier models skipped public evals, notes researcher — jessi_cata · 2026-11-06
- webAI launches open 8B model and Mac app for on-device collaborative AI agents — omarsar0 · 2026-11-06
- OpenAI's newest model claimed to have solved 100+ long-standing open math problems in two weeks — felpix_ · 2026-11-06
- Qwen3.8-27B hits 381 tok/s on a single RTX 3090 via inference software alone — Redo · 2026-11-06
- GPTs believe in the Dark Forest more than humans: the 'philosophical shame' of refusals — repligate · 2026-11-06