A Simpler and Better Debate Protocol for Scalable Oversight

New research challenges the superposition theory of polysemanticity, suggesting neural networks are more interpretable than previously believed through recursive generalization patterns.

Nick Cammarata at Goodfire AI claims neural networks are 'clean' with 90% of Qwen-32B neurons having single functions, disproving polysemanticity and enabling direct study without model unfolding.

Nick Cammarata argues neural networks are already clean and interpretable, challenging previous pessimism about mechanistic interpretability.

Researchers introduce OmniVerifier-C1, a decomposed graded verifier that provides continuous scoring instead of binary judgments for multimodal verification tasks.

Researchers introduce T2-z, a method for test-time optimization of environment latents in robotic world models that improves prediction accuracy and action success rates across simulations and real-world tasks.