Decoupled Marginal Sharpening for Training-Free Inference-Time Scaling
cs.LG
Submitted: 2026-10-14
Updated: 2026-10-14
Comments: 17 pages, 3 figures, 6 tables
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Self-Consistency via Marginal Sharpening
- Power-SMC: Low-Latency Sequence-Level Power Sampling for Training-Free LLM Reasoning
- Ensembling Language Models with Sequential Monte Carlo
- Evaluating Large Language Models Trained on Code
- Score $\times$ Decoder: A Unified View of Unsupervised Inference-Time Scaling for Hallucination Mitigation
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Measuring Mathematical Problem Solving With the MATH Dataset
- Scalable Power Sampling: Unlocking Efficient, Training-Free Reasoning for LLMs via Distribution Sharpening
- Reasoning with Sampling: Your Base Model is Smarter Than You Think
- Sampling for Quality: Training-Free Reward-Guided LLM Decoding via Sequential Monte Carlo
- The Model Knows, the Decoder Finds: Future Value Guided Particle Power Sampling
- GPQA: A Graduate-Level Google-Proof Q&A Benchmark
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- Inference-Time Alignment of Diffusion Models via Trust-Region Iterative Twisted Sequential Monte Carlo
- Self-Consistency Improves Chain of Thought Reasoning in Language Models
- Qwen3 Technical Report
- When are likely answers right? On Sequence Probability and Correctness in LLMs
- Embarrassingly Simple Self-Distillation Improves Code Generation
- Probabilistic Inference in Language Models via Twisted Sequential Monte Carlo
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