inference efficiency

AutoTTS lets LLMs discover test-time scaling

AutoTTS turns test-time scaling into an environment search problem, letting LLMs discover cheaper reasoning strategies automatically.