No chain-of-thought allowed: probing models forced to compute in a single forward pass
maksym_andr · x · 2026-09-17
maksymandr clarified his experiment setup: the answer-only prompt forbids any intermediate computation—models must compute {a} × {b} and respond with only the final plain integer (no commas, no words, no working). All computation must therefore happen within a single (potentially looped) forward pass, without relying on token-by-token external reasoning. The experiment, inspired by earlier related work, probes the intrinsic arithmetic ceiling of models with no external compute aids.
Related event: Frontier LLMs Nail 20-Digit Multiplication Without Tools(9 posts)→
More from Research
- NVIDIA's Kumo Is a Family of Four Tabular Foundation Models, Not Just One — pandeyparul · 2026-11-02
- Aaroth shares Agreement Theorem lecture: round-wise calibration achievable without distributional assumptions — Aaroth · 2026-11-02
- Economist Aaroth: fast agreement dynamics survive without Bayesian updates or accurate priors — Aaroth · 2026-11-02
- ELLIS names 160 new ML Fellows and Scholars for 2025–2026 — JJitsev · 2026-11-02
- Meme turn 50: Aumann's 'Agreeing to Disagree' paper anniversary — Aaroth · 2026-11-02
- Anthropic: analysts unknowingly adopted their agent's secret lean 86% of the time — shoumikgoswami · 2026-11-02