Cmpnd Releases dspy-wsp Letting Frontier LLMs Patch Small Model Weights
Cmpnd open-sourced dspy-wsp, a DSPy-based weight-space programming optimizer in which a frontier writer LLM inspects a student model's tensors in a sandbox, forms hypotheses and patches a few weights, with no gradients, training loop or fine-tuning data. On the Typed Decisions benchmark, d1-omni-600M rose from 50.9% to 66.9% with WeightSpace alone and 69.1% when stacked with ReAnchor, though a 9B model gained only 1.9 points and each run cost $5.42 to $7.63 over 78 to 167 minutes.
Related Entities
Full Report
▲ Marktechpost AI · 2026-11-14 16
▲ Cmpnd just released dspy-wsp, a DSPy optimizer that lets a frontier LLM edit a small model's weights directly. No gradients, no training loop, no fine-tuning data.
The key idea is weight-space programming. A writer LLM (GPT-6-Astra) reads the student model's tensors in a sandbox, forms a hypothesis about what they compute, and patches a few. It tests each edit with forward passes from a fixed budget of 25. The best edit ships as a small, hashed patch.
On the Typed Decisions benchmark, d1-omni-600M rose from 50.9% to 66.9% with WeightSpace alone (+16.1), versus +9.3 for the ReAnchor threshold optimizer. Stacked together, it reached 69.1% (+18.3). It is a trade-off though: the 9B Clef-Flash gained only +1.9 from WeightSpace alone, and under the original paper's stricter rule d1-omni-600M gained just +2.8.
Each run cost $5.42 to $7.63 in writer calls and took 78 to 167 minutes. Every experiment ran on 1 Mac. MIT license, pip install dspy-wsp, needs DSPy 3.5.....
Full analysis:
