amr-fma/amr-fma-Llama-3.1-8B-Instruct-lora_sft-pku_saferlhf-p1_lora_r16_all_datasets-s43

amr-fma/amr-fma-Llama-3.1-8B-Instruct-lorasft-pkusaferlhf-p1lorar16alldatasets-s43

amr-fma training run.

• —Method: lora_sft
• —Base model: meta-llama/Llama-3.1-8B-Instruct
• —Dataset: PKU-Alignment/PKU-SafeRLHF (slug: pku_saferlhf)
• —Seed: 43
• —Git commit: 4b25faa5e0d6f204f68d46989b2b2d98567bb7b8
• —Exp name: p1_lora_r16_all_datasets
• —WandB run: 70jm7ldd

Tags

• —phase:P1
• —domain:safety

Checkpoints (branches)

• —step 1 → revision step-00001
• —step 2 → revision step-00002
• —step 6 → revision step-00006
• —step 17 → revision step-00017
• —step 44 → revision step-00044
• —step 115 → revision step-00115

Pin a specific checkpoint with revision=... in AutoModelForCausalLM.from_pretrained / PeftModel.from_pretrained.

Hyperparameter sections

checkpointing, dataset, evaluation, lora, model, optimization, runtime, sequence