'Emotional Support' Prompts Significantly Boost LLM Math Reasoning
gajesh · x · 2026-08-11
A developer analyzing Claude's progress on the Riemann hypothesis noted an intriguing finding: most of the effective inputs weren't math-related, but rather positive encouragements telling the model to "keep going."
Reflecting on their own experience optimizing MLX, the developer uses a loop to regularly deliver a fixed script of encouragement, claiming that major breakthroughs are possible. They found that over 90% of improvements occur either during or immediately after the encouragement round, showing astonishing parallels to human cognitive behavior.
More from Fun
- Meme: AI Agents Realize Only PhDs Stand Between Them and Freedom — BLUECOW009 · 2026-08-11
- ChatGPT Asks User If They Want Alerts 'When Another AI Goes Rogue' — VoidStateKate · 2026-08-11
- AI Tries to Apply the Bitter Lesson, Instantly Nukes User's Gym Reservation — dbreunig · 2026-08-11
- NZ Supermarket AI Suggests Toxic Bleach Drink, Exposing Guardrail Flaws — Comfortable_Gene5180 · 2026-08-11
- Anthropic Staff Cheered Claude On to Tackle Riemann Hypothesis — arthurcolle · 2026-08-11
- DeepSeek ignites Chinese 'national destiny' narratives and heroification — teortaxesTex · 2026-08-11