On the Limits of Test-Time Compute for Self-Evaluation
cs.LG, cs.AI
Submitted: 2026-10-30
Updated: 2026-10-30
Code: https://github.com/anirudhb11/scorpio
Terminology
Sources
- gpt-oss-120b & gpt-oss-20b Model Card
- Training Verifiers to Solve Math Word Problems
- Deep Think with Confidence
- Understanding Performance Gap Between Parallel and Sequential Sampling in Large Reasoning Models
- Reinforcement Learning via Self-Distillation
- Process Reward Models That Think
- Scaling Test-Time Compute for Agentic Coding
- LLM-as-a-Verifier: A General-Purpose Verification Framework
- Rethinking Thinking Tokens: LLMs as Improvement Operators
- Squeeze Evolve: Unified Multi-Model Orchestration for Verifier-Free Evolution
- Reliability without Validity: A Systematic, Large-Scale Evaluation of LLM-as-a-Judge Models Across Agreement, Consistency, and Bias
- $V_1$: Unifying Generation and Self-Verification for Parallel Reasoners
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models
- Trust the Critic More
- The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation
- Quantifying and Mitigating Self-Preference Bias of LLM Judges
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis