AI Briefing#self-distillation
The latest AI and developer news about #self-distillation, with the original source and a short summary.
The latest AI and developer news about #self-distillation, with the original source and a short summary.
A study on GPT-OSS-20B shows that a fixed generic feedback line yields the same ~3-point performance gain as self-written critiques in multi-step financial reasoning.
The method uses self-generated question-answer pairs and reasoning traces to boost performance across text and multimodal domains while retaining capabilities in untrained areas.
RISED uses LLM-generated rubrics to guide data selection and self-distillation, achieving the highest mean pass rate across diverse interactive environments.
The framework improved GenEval scores from 0.747 to 0.808 on Qwen-image-2512 by using the model's own critiques as training signals.
GLM 5.2 lowered the Tool-call failure rate by 21.2% through combined training using RFT on real user session data and Hint-guided Self-distillation.
Applying the model grafting technique, which converts existing models into a causal encoder-decoder architecture, increased Qwen3.5-4B's prompt processing speed by up to 3.7x.
LLMs can significantly improve their code generation ability using only their own outputs, without needing a verifier or separate training.
A new technique is proposed that compresses narrative portions of the reasoning process while preserving only the computation/verification segments, reducing tokens while maintaining performance.
Self-Distillation Fine-Tuning reduced forgetting in demonstration-based learning.
NTM aims for both 4-step generation and exact likelihood by training the reverse step as a normalizing flow.
A weight-sharing looped Transformer improves the efficiency of image and video generation.
A self-distillation (SSD) technique that uses only the model's own outputs, without a separate teacher model or reinforcement learning, significantly improved LLMs' code generation performance.