arxiv
PublishedJune 2, 2026 at 4:00 AM
—neutral
The Shape of Wisdom: Decision Trajectories in Language Models
Publisher summary· verbatim
arXiv:2606.01202v1 Announce Type: new Abstract: Language models do not simply choose an answer at the output layer. In a 9,000-trajectory MMLU study across Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct, and Mistral-7B-Instruct-v0.3, the score of the answer moves across depth in structured ways. We desc
Stay posted· Newsletter
A 5-min weekly brief — top movers, price watch, story of the week.
Discussion
No replies yet. Be first.
Related coverage
More from ARXIV
arxivFederatedSkill: Federated Learning for Agentic Skill Evolution13harxivToward a Modular Architecture for Embedded AI Agent Systems at the Edge13harxivA Graph Foundation Model with Spectral Parsing and Prototype-Guided Spatial Propagation13harxivAnomalies in Multivariate Time Series Benchmarks Are Mostly Univariate13hThe Bubble Brief
WEEKLYRead AI insights every Tuesday — top movers, new releases, story of the week.
Originally published on arxiv ↗