Teaching LLMs to Say 'I Need More Info' Instead of Guessing: ACA-RL Trains Models on Incomplete Problems

Teaching LLMs to Say 'I Need More Info' Instead of Guessing: ACA-RL Trains Models on Incomplete Problems | BedrockNews