Weight-Space Programming can improve reasoning performance through small, localized edits to neural network parameters, but its editing reliability and cross-task generalization remain uneven.
Weight-Space Programming (WSP) enables reasoning agents to directly inspect and modify neural network parameters, offering an alternative to gradient-based model adaptation. However, its capabilities, cross-task generalization, and practical limitations remain insufficiently understood. We systematically investigate WSP through 16 editing trajectories across seven benchmarks and two model scales (Qwen3.5-0.8B and Qwen3.5-9B). Our findings reveal three key insights. First, WSP can achieve substantial performance gains through highly localized interventions. On full held-out evaluations, accuracy improves by up to 29.45, 4.42, 36.72, and 17.11 percentage points on MATH-500, GPQA Diamond,…
Weight-Space ProgrammingWSPreasoning agentsneural network parametersgradient-based model adaptationediting trajectoriescross-task generalizationlocalized interventionsheld-out evaluationscheckpoints
- Upvotes
- 6
- Authors
- 8
- Published
- Nov 6, 2026
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.