model-layers
Qwen3-30B-A3B-NVFP4-2-layers-modeloptlens-loss-minimality_lr_1e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_1_num_layers_6lens-loss-minimality_lr_2e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_1_num_layers_6lens-loss-minimality_lr_2e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_1_num_layers_32lens-loss-minimality_lr_2e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_1_num_layers_24lens-loss-minimality-l2_lr_2e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_1_num_layers_6adversarial_loss_lr_1e-5_model_meta-llama_Llama-2-7b-chat-hf_batch_4_epoch_4_num_layers_6toxicity-reward-model-output-max-margin-5-seed-200-unfrozen-layers-0-pythia-70m