frontier models
Keep the Best Thoughts, but Not Copies of Them
A pruning rule optimizes the chance that at least one surviving reasoning path succeeds, balancing quality against redundant trajectories.
Summary
A pruning rule optimizes the chance that at least one surviving reasoning path succeeds, balancing quality against redundant trajectories.
Test-time reasoning can generate many partial solutions, then spend most of its compute finishing only a few. Pure quality scoring risks keeping near-duplicates; pure diversity can preserve weak paths. The proposed selector directly optimizes pass-at-k for the surviving pool, rewarding promising prefixes while penalizing overlap in the final answers they are likely to reach. On eight open-weight reasoning models, it outperformed the quality-only and diversity-oriented pruning baselines, with the largest gains when a moderate number of continuations survived. The result concerns the tested reasoning tasks and estimators, not a universal inference policy.
Why it matters
A pruning rule optimizes the chance that at least one surviving reasoning path succeeds, balancing quality against redundant trajectories.
Limits and context
- Test-time reasoning can generate many partial solutions, then spend most of its compute finishing only a few.
- On eight open-weight reasoning models, it outperformed the quality-only and diversity-oriented pruning baselines, with the largest gains when a moderate number of continuations survived.
- The result concerns the tested reasoning tasks and estimators, not a universal inference policy.
Key claims
A pruning rule optimizes the chance that at least one surviving reasoning path succeeds, balancing quality against redundant trajectories.
Qualification: Test-time reasoning can generate many partial solutions, then spend most of its compute finishing only a few.
Evidence: source-2026-10-23-006
Sources
- arXiv preprint 2610.31667, original v1arXiv · primary research
Corrections
No corrections have been recorded for this story.