SEP 8, 2026 · PREPRINT
ThinkPrior: Zero-Rollout Difficulty Priors for Cold-Start Prompt Selection in RLVR
arXiv
Unreviewed computational method study showing early-stage algorithmic improvement in rollout efficiency during cold-start prompt selection, with no difference in final task accuracy and limited scope (single model, 250-prompt pool).
Reported
Silent-group reductionmore than halved
Wasted rollout reduction through…nearly a fifth
Generated rollout reduction (Thin…10.6%