This model is a fine-tuned version of W-61/llama-3-8b-base-sft-hh-harmless-8xh200 on the Anthropic/hh-rlhf dataset. It achieves the following results on the evaluation set:

Model description

More information needed

Intended uses & limitations

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 5e-07
train_batch_size: 16
eval_batch_size: 16
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 128
total_eval_batch_size: 128
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 1

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/chosen	Logps/rejected	Logps/ref Chosen	Logps/ref Rejected	Logits/chosen	Logits/rejected	Kl/p Epsilon Steps	Kl/n Epsilon Steps
0.6773	0.3030	100	0.6737	-0.1469	-0.1956	0.6085	0.0487	-91.4063	-106.5723	-73.9652	-83.1798	0.4327	0.4096	0.5960	0.4030
0.6222	0.6061	200	0.6386	-0.3851	-0.5821	0.6572	0.1970	-131.2605	-170.4228	-73.9652	-83.1798	0.6604	0.6363	0.6351	0.3640
0.6236	0.9091	300	0.6288	-0.4104	-0.6458	0.6691	0.2353	-157.0541	-214.8666	-73.9652	-83.1798	0.7149	0.6886	0.6448	0.3534