Dolly-15k

Researchers introduced DataPrep-Bench, the first unified benchmark for evaluating how well large language models (LLMs) and agents prepare training data, covering both data construction and data quali…

A new study on Llama 3.1 8B finds that helpfulness post-training (SFT and GRPO) significantly degrades animal compassion values compared to coding-domain post-training, with a 35.7% vs. 65.2% gap on t…