cby-pku/aligner

Issues

• Warm-up训练

  #7 opened 15 days ago
  5

• question about the paper

  #6 opened 4 months ago
  1

• The code is training Aligner or using the Aligner to train the strong-model?

  #5 opened 6 months ago
  1

• 训练对齐器，显存溢出

  #4 opened 6 months ago
  1

• WANDB_API_KEY是什么意思，在哪里找到

  #3 opened 6 months ago
  1

• How to evaluate on Beavertails and HarmfulQA

  #2 opened 8 months ago
  1

• Aligner-7B improves 11 different LLMs by 21.9% in helpfulness and 23.8% in harmlessness on average (GPT-4 by 17.5% and 26.9%).

  #1 opened 8 months ago
  1