Can Weaker Models Replicate Frontier Discoveries with Hints? Exploring LLM Basins of Attraction
danshipper · x · 2026-08-03
AI researcher Dan Shipper tasked GPT-5.6 with a mathematical challenge before a flight: given a hint involving algebraic number theory, can it reproduce Astra's proof of Erdős's planar unit distance conjecture?
He uses this to propose a broader theory: weaker models can often reproduce frontier-model discoveries if provided with the right conceptual hints. A stronger model's advantage is that it can start farther from the answer, effectively possessing a larger basin of attraction around the correct solution. He suggests this approach could generalize into a novel benchmark for evaluating model generalization and reasoning outside their training data.
Related event: Prompt Distance: Can Weaker Models Reproduce Frontier Proofs?(6 posts)→
More from AGI Musings
- Gary Marcus: AI firms self-testing safety is like tobacco companies grading themselves — GaryMarcus · 2026-09-18
- In 1881 NYT Cited Academics Warning Telegraphy Could End the World, Echoing Today's AI Doom — arampell · 2026-09-18
- OpenAI's Noam Brown: Air-Gapping May Not Stop a Misaligned AI, Bar Must Be 'Very, Very High' — deanwball · 2026-09-18
- OpenAI exec: GPUs hit 7-40 IQ points per watt vs human's 5, a milestone we 'zoomed past' — GregCook2011 · 2026-09-18
- AI is scrambling wages: blue-collar jobs hit $200K-$500K as white-collar work dries up — bindureddy · 2026-09-18
- Gary Marcus: Judea Pearl's causality challenges remain unsolved in the LLM paradigm — GaryMarcus · 2026-09-18