Research · AWS ML Blog ·

Exploring self-distilled reasoning for supervised fine-tuning with Amazon Nova

Exploring self-distilled reasoning for supervised fine-tuning with Amazon Nova

AWS presents Self-Distilled Reasoning (SDR), a method for generating reasoning tokens when supervised fine-tuning datasets lack reasoning traces. The post examines reasoning suppression, evaluates SDR on three benchmarks, and offers implementation recommendations for Amazon Nova customization.

Read the full story at AWS ML Blog →