Mode-Conditioning Unlocks Superior Test-Time Scaling

N/A

Preprint Mode-Conditioning Unlocks Superior Test-Time Scaling Chen Henry Wu, Sachin Goyal, Aditi Raghunathan Carnegie Mellon University # {chenwu2,sachingo,aditirag}@cs.cmu.edu Ð https://github.com/AR-FORUM/modc ƒ https://ar-forum.github.io/mod-c/ arXiv:2512.01127v1 [cs.LG] 30 Nov 2025…

saved by

• Ashvin Nair

related reading

• [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
• [2501.19393] s1: Simple test-time scalingarxiv.org
• Weight Ensembling Improves Reasoning in Language Modelsarxiv.org
• [2505.17373] Value-Guided Search for Efficient Chain-of-Thought Reasoningarxiv.org
• [2604.22082] Removing Sandbagging in LLMs by Training with Weak Supervisionarxiv.org
• 2409.02908arxiv.org
• Generalization Dynamics of LM Pre-training — Jiaxin Wenjiaxin-wen.github.io
• [2509.14786] Pre-training under infinite computearxiv.org
• Optimizing LLM Test-Time Compute Involves Solving a Meta-RL Problem – Machine Learning Blog | ML@CMU | Carnegie Mellon Universityblog.ml.cmu.edu
• [2603.07267] How to Steal Reasoning Without Reasoning Tracesarxiv.org
• [2605.06546] Efficient Pre-Training with Token Superpositionarxiv.org
• IsoCompute Playbook: Optimally Scaling Sampling Compute for RL Training of LLMscompute-optimal-rl-llm-scaling.github.io