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