Early Signals, Strong Decisions: Prefix-Guided Sampling for Parallel Test-Time Scaling
cs.AI
Submitted: 2026-11-11
Updated: 2026-11-14
Code: https://github.com/renjie3/PreG
Terminology
Sources
- OpenAI o1 System Card
- GPT-4 Technical Report
- Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- Training Verifiers to Solve Math Word Problems
- Parallel Test-Time Scaling with Multi-Sequence Verifiers
- Deep Think with Confidence
- Efficiently Scaling LLM Reasoning with Certaindex
- PiCSAR: Probabilistic Confidence Selection And Ranking for Reasoning Chains
- Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokens
- TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
- MatryoshkaThinking: Recursive Test-Time Scaling Enables Efficient Reasoning
- Sampling-Efficient Test-Time Scaling: Self-Estimating the Best-of-N Sampling in Early Decoding
- CoRefine: Confidence-Guided Self-Refinement for Adaptive Test-Time Compute
- Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs
- Believe Your Model: Distribution-Guided Confidence Calibration
- The Sequential Edge: Inverse-Entropy Voting Beats Parallel Self-Consistency at Matched Compute
- CGES: Confidence-Guided Early Stopping for Efficient and Accurate Self-Consistency
- Qwen3 Technical Report
- Llama-Nemotron: Efficient Reasoning Models
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents