Researcher questions whether multi-question decider models are architecturally misused
antoine_chaffin · x · 2026-10-19
antoinechaffin shares his thinking on attention schemas for decider models, arguing we may not even be using these models properly: optimizing schemas for requests that should never have been forged that way may not be optimal. He questions whether any use case truly requires multiple questions at once — most could be single questions with multiple choices, or pointwise multi-label outputs with proper thresholds instead of a softmax — and contends that even multi-question setups could be served without state-question cross-attention if the state is constructed properly.
Related event: Ex-OpenAI Researcher Questions Design of Multi-Question Decision Models(3 posts)→
More from Models
- Rumor: Anthropic training a model matching OpenAI's internal one; big launches in December — imjustnewatai · 2026-10-23
- AI researcher: Claude finally finished tedious research tasks I gave up on years ago — MikePFrank · 2026-10-23
- Dream Labs emerges from stealth with Dream-1 robot foundation model at $3/hour — scott_e_reed · 2026-10-23
- Claude Models on Google Cloud (Anthropic managed) is now generally available — claudeai · 2026-10-23
- Beam vs GLM 5.3 Abliterated: 10-round attack/defense test shows 96% of vulns patched — bigblueboo · 2026-10-23
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23