BWhat changed
Adaptively chose between fresh candidate expansion and feedback-based refinement of existing solution nodes, balancing broad sampling and deep search during language-model inference.
Scientific scope
Adaptively chose between fresh candidate expansion and feedback-based refinement of existing solution nodes, balancing broad sampling and deep search during language-model inference. MCTS, repeated sampling, feedback and bandit exploration are inherited. Original March report controls method; later multi-model ARCAGI2 extension is not backdated. Search quality depends on feedback correctness and inference budget, not guaranteed reasoning success.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. MCTS, repeated sampling, feedback and bandit exploration are inherited. Original March report controls method; later multi-model ARCAGI2 extension is not backdated. Search quality depends on feedback correctness and inference budget, not guaranteed reasoning success.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search
- Misaki et al.: feedback-adaptive width/depth inference search — author/institutional project
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction