SIMPLEMIX: Frustratingly Simple Mixing of Off- and On-policy Data in Language Model Preference Learning

Documenti analoghi

• Verifiable by Design: Aligning Language Models to Quote from Pre-Training Data
  di: Zhang, Jingyu, et al.
  Pubblicazione: (2024)
• The Flaw of Averages: Quantifying Uniformity of Performance on Benchmarks
  di: Uzunoglu, Arda, et al.
  Pubblicazione: (2025)
• Error Norm Truncation: Robust Training in the Presence of Data Noise for Text Generation Models
  di: Li, Tianjian, et al.
  Pubblicazione: (2023)
• Benchmarking Language Model Creativity: A Case Study on Code Generation
  di: Lu, Yining, et al.
  Pubblicazione: (2024)
• The Translation Barrier Hypothesis: Multilingual Generation with Large Language Models Suffers from Implicit Translation Failure
  di: Bafna, Niyati, et al.
  Pubblicazione: (2025)