Atlas · GenAI 2026
Test-Time Compute Scaling
Trading extra inference-time compute for accuracy — longer reasoning chains, best-of-N sampling, self-consistency, verifier-guided search.
conceptPeak: 2025Test-Time ComputeAI consensus: 0/3
Prerequisites
It manipulates how tokens are sampled/searched at inference.
- mediumIn-Context Learning
Chains of thought are prompted in-context.
Recommended reference
Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Parameters — arXiv 2408.03314