LLM News

Researchers have identified and solved a critical scaling problem in LLM-based theorem prediction called Structural Drift, where in-context learning performance collapses as reasoning depth increases. Using Theorem Precedence Graphs to encode topological dependencies, they achieved 89.29% accuracy on the FormalGeo7k benchmark—matching state-of-the-art supervised approaches without any gradient-based training.

A new study challenges the assumption that chain-of-thought reasoning translates directly to biological domains. Researchers found that text-only reasoning for protein function prediction produces superficial patterns rather than new biological knowledge. A tool-augmented agent called PFUA achieves 103% average performance improvement by integrating domain-specific tools for verifiable intermediate evidence.