Cmpnd has released dspy-wsp, a new optimizer for the DSPy framework that enables weight-space programming. This technique allows a large language model to directly edit the weights of another model without traditional training loops or gradients. In tests, using GPT-6-Astra to modify weights for a few dollars significantly improved decision models, with gains ranging from 1.9 to over 18 points on specific benchmarks. AI
IMPACT Enables gradient-free model editing, potentially simplifying and reducing the cost of model improvement.
RANK_REASON Release of a new optimizer tool for an existing framework.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
