Socratic-Zero co-evolves agents to generate reasoning data
The approach uses three agents and 100 initial questions to produce reasoning data; the paper reports a 20.2-point advantage across seven mathematical reasoning benchmarks.
Socratic-Zero combines the Teacher, Solver, and Generator agents to create reasoning data from 100 initial questions. The paper reports that the Socratic-Solver-8B model outperforms previous data-synthesis methods by 20.2 points across seven mathematical reasoning benchmarks.
The proposal describes generating reasoning curricula without human-labeled data. If you use AI to study or adapt this kind of method, avoid putting personal data or internal information in prompts, and follow your organization’s data-handling rules.