metadata
license: mit
base_model: Qwen/Qwen3-4B
Test-time Calibration Learning (TTCL)
TTCL is performed on ChineseSimpleQA benchmark based on RLCR-Qwen3-4B-DAPO-Math14K.
license: mit
base_model: Qwen/Qwen3-4B
TTCL is performed on ChineseSimpleQA benchmark based on RLCR-Qwen3-4B-DAPO-Math14K.