daixuancheng/zero_7b_base_useTokenLoss_clipHigh_KLcoeff0_step200
Text Generation • 8B • Updated • 5
Daixuan Cheng
daixuancheng
AI & ML interests
I work on LLMs across pre-training, post-training, and agents.
Recent Activity
updated a model 2 days ago
daixuancheng/Qwen3-4B-Instruct-2507-LLM-in-Sandbox-RL updated a dataset 2 days ago
daixuancheng/llm-in-sandbox-bench updated a dataset 2 days ago
daixuancheng/llm-in-sandbox-rlOrganizations
daixuancheng/ppo_sample8_critic-warm10-lr2e-6_step200_crtic
Text Generation • 8B • Updated • 6
daixuancheng/ppo_sac_static0.1_constrainbyadv_step-200_critic
Text Generation • 8B • Updated • 7
daixuancheng/sac-init0.4_qwen-math-7b_constrainbyAdv_yesSuffix_step200
Text Generation • 8B • Updated • 6
daixuancheng/zero_7b_base_useTokenLoss_clipHigh_KLcoeff0_step180
Text Generation • 8B • Updated • 5
daixuancheng/rerun_sac-init0.1_qwen-math-7b_constrainbyAdv_step240
Text Generation • 8B • Updated • 5
daixuancheng/fix-entropy-1e-3_train_math_global_step_200
Text Generation • 8B • Updated • 5
daixuancheng/ppo_sample8_critic-warm10-lr2e-6_step200_actor
Text Generation • 8B • Updated • 6
daixuancheng/ppo_sac_static0.1_constrainbyadv_step-200_actor
Text Generation • 8B • Updated • 6
daixuancheng/rerun_qwen-math-7b_noSuffix_base_step280
Text Generation • 8B • Updated • 6
daixuancheng/fix-entropy-1e-3_train_math_global_step_180
Text Generation • 8B • Updated • 5
daixuancheng/rerun_sac-init0.1_qwen-math-7b_constrainbyAdv_step260
Text Generation • 8B • Updated • 4
daixuancheng/rerun_qwen-math-7b_noSuffix_base_step260
Text Generation • 8B • Updated • 8
daixuancheng/ppo_sample8_critic-warm10-lr2e-6_step180_crtic
Text Generation • 8B • Updated • 4
daixuancheng/ppo_sample8_critic-warm10-lr2e-6_step180_actor
Text Generation • 8B • Updated • 7
daixuancheng/zero_qwen-math-7b_base_allDapo_mathVerify_yesSuffix_step220
Text Generation • 8B • Updated • 6
daixuancheng/sac_static0.1_constrainbyAdv_step180
Text Generation • 8B • Updated • 6
daixuancheng/sac-init0.4_qwen-math-7b_constrainbyAdv_yesSuffix_step220
Text Generation • 8B • Updated • 7
daixuancheng/zero_7b_base_useTokenLoss_clipHigh_KLcoeff0_step160
Text Generation • 8B • Updated • 6
daixuancheng/sac_static0.4_constrainbyAdv_step160_bak
Text Generation • 8B • Updated • 6
daixuancheng/ppo_sac_static0.1_constrainbyadv_step-180_critic
Text Generation • 8B • Updated • 6
daixuancheng/ppo_sac_static0.1_constrainbyadv_step-180_actor
Text Generation • 8B • Updated • 7
daixuancheng/fix-entropy-1e-3_train_math_global_step_160
Text Generation • 8B • Updated • 13
daixuancheng/fix-entropy-1e-3_train_math_global_step_140
Text Generation • 8B • Updated • 6