Quality-Diversity Stress Tests for Process Reward Models:What Archive Coverage Can and Cannot Certify
Ibne Farabi Shihab, Fariya Afrin
cs.LG
Submitted: 2026-08-08
Updated: 2026-08-11
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Concrete Problems in AI Safety
- RainbowPlus: Enhancing Adversarial Prompt Generation via Evolutionary Quality-Diversity Search
- Adversarial Training for Process Reward Models
- Illuminating search spaces by mapping elites
- EST-PRM: Stress-Testing Process Reward Models Before They Become Load-Bearing
- PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models
- Solving math word problems with process- and outcome-based feedback
- Quality-Diversity Red-Teaming: Automated Generation of High-Quality and Diverse Attackers for Large Language Models
- The Lessons of Developing Process Reward Models in Mathematical Reasoning
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks