Tracked topic
…DataCOPE derives verifier signals from the exploration trajectories and uses them to characterize relative quality or aggreement among trajectories. It iteratively coordinates a Data-Analytic Agent for trajectory generation, an Unsupervised Verifier…
…Yang Tian , , , , , , , Abstract Privileged Bayesian Self-Distillation enables fine-grained credit assignment in long-horizon tasks by converting sparse outcome rewards into calibrated turn-level signals through Bayesian evidence scoring and autoregressive…
…A subsumption test shows that phi_first is moderately to strongly correlated with semantic agreement , and combining the two signals yields only a small AUROC improvement over phi_first alone. These results…
…Sustaining Perception for Deep Generation in LVLMs Published on May 1 Submitted by Siyuan Huang on May 5 Authors: , , , , , , , , Abstract Large vision-language models suffer from visual signal dilution during long sequence…
…It leverages the base model 's knowledge without any external signal. Our approach pairs instruction-following cues extracted from the frozen model with cycle-consistency for structure preservation. To make this tractable…
…Existing approaches to such process credit assignment either depend on separate external process reward models that introduce additional consumption, or tree-based structural rollout that merely redistributes the outcome signal while constraining…
…General-purpose compression schemes based on scalar quantization or resolution reduction are broadly applicable but fail to exploit inherent signal redundancies, resulting in suboptimal rate-distortion performance . Recent generative neural codec s…
…would love to see an ablation where you remove one expert or replace them with a shared critic to see which signal drives the improvement more. also, since alignment is on-policy…
…Generated by Qwen/Qwen2.5-Coder-32B-Instruct We present ChildVox, a novel benchmark for characterizing the diverse acoustic signals through which children communicate. Specifically, ChildVox follows the full developmental trajectory from…
…First, Stepwise Reward Centering subtracts expected rewards to neutralize length-dependent bias , ensuring that path extension yields zero expected gradient signal. Second, Position-Specific Advantage Estimation leverages the reward decomposition structure to…