task · first seen 5 Oct 2026
process reward verification
Also: step-level verification
Checking the correctness of individual reasoning steps in a model-generated solution.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Can Language Models Learn to Reject Their Own Bad Reasoning Steps?Papers · 5 Oct 2026 · Siheng Xiong, Xiaoze Liu, Yiqiao Jin, Xiaoqian Wang, Jing Gao
- Attention-Graph Curvature for the Verification of Mathematical ReasoningPapers · 12 Oct 2026 · Kanghyeon Kim, Johannes F. Lutzeyer, Yanzhu Guo
- Where Search Agents Go Off the Rails: Graph-Grounded Step-Level VerificationPapers · 4 Nov 2026 · Hyeseo Jeon, Seonho Yoo, Jay-Yoon Lee
Latest sources
- Where Search Agents Go Off the Rails: Graph-Grounded Step-Level VerificationPapers · 4 Nov 2026 · Hyeseo Jeon, Seonho Yoo, Jay-Yoon Lee
- Attention-Graph Curvature for the Verification of Mathematical ReasoningPapers · 12 Oct 2026 · Kanghyeon Kim, Johannes F. Lutzeyer, Yanzhu Guo
- Can Language Models Learn to Reject Their Own Bad Reasoning Steps?Papers · 5 Oct 2026 · Siheng Xiong, Xiaoze Liu, Yiqiao Jin, Xiaoqian Wang, Jing Gao