Best-of-N

A new arXiv paper (2610.09250v1) proposes a sample-only framework for evaluating and selecting Best-of-N (BoN) inference-time alignment policies without access to response likelihoods, which standard …

Researchers propose Lagrangian Reward Augmentation (LARA), a framework for inference-time alignment of language models that enforces safety constraints by dualizing a constrained optimization problem …

Researchers introduce Temporal Backtracking Search (TBS), a test-time scaling method for generative video reasoning that searches over the temporal axis rather than denoising steps. TBS transforms vid…