TY  - RPRT
TI  - Optimizing Pretraining Data Mixtures with LLM-Estimated Utility
AU  - William Held
AU  - Bhargavi Paranjape
AU  - Punit Singh Koura
AU  - Mike Lewis
AU  - Frank Zhang
AU  - Todor Mihaylov
PY  - 2025
UR  - https://arxiv.org/abs/2501.11747
ID  - 2501.11747
ER  -