base_model_name_1 : ./logs_morlhf/rlhf_faithful/batch_300
base_model_name_2 : ./logs_morlhf/rlhf_summary/batch_300
base_model_name_3 : 
reward_peft_path1 : reward_models/CogComp/bart-faithful-summary-detector
reward_peft_path2 : reward_models/Tristan/gpt2_reward_summarization
tokenier_name : meta-llama/Llama-2-7b-hf
