Trained ExpRL checkpoints. Paper link: https://arxiv.org/abs/2606.17024
Violet Xiang PRO
violetxi
AI & ML interests
None yet
Recent Activity
upvoted a paper about 2 hours ago
ExpRL: Exploratory RL for LLM Mid-Training updated a collection about 2 hours ago
ExpRL submitted a paper about 2 hours ago
ExpRL: Exploratory RL for LLM Mid-TrainingOrganizations
models 32
violetxi/ExpRL-Outcome-Qwen3-8B
8B • Updated
violetxi/ExpRL-Process-Qwen3-4B-Instruct
4B • Updated
violetxi/ExpRL-Outcome-Qwen3-4B-Instruct
4B • Updated
violetxi/sft-clbench-poker-wm-sft-poker-qwen3-5-4b-future-summary
5B • Updated • 15
violetxi/sft-clbench-poker-wm-sft-poker-qwen3-5-4b-next-state
5B • Updated • 9
violetxi/sft-clbench-poker-wm-sft-poker-qwen3-5-4b
5B • Updated • 14
violetxi/teacher_tooluse_grpo_kl-1
4B • Updated • 258
violetxi/opsd_sciknoweval_chemistry_qwen3-4b_frozen_teacher_forward_kl_bs256_lr5e-6
4B • Updated • 16
violetxi/opd_tooluse_qwen3-4b_trained_teacher_forward_kl_bs256_lr5e-6
4B • Updated • 2
violetxi/opd_tooluse_qwen3-4b_trained_teacher_forward_kl_bs256_lr5e-6_antileak
4B • Updated • 2
datasets 161
violetxi/clbench-poker-halftest-traj-gemini-future-summary
Viewer • Updated • 1.76k • 58
violetxi/clbench-poker-halftest-traj-qwen3-4b-future-summary
Viewer • Updated • 55 • 22
violetxi/clbench-poker-halftest-traj-gemini-next-state
Viewer • Updated • 1.76k • 54
violetxi/clbench-poker-halftest-traj-qwen3-4b-next-state
Viewer • Updated • 55 • 21
violetxi/clbench-poker-halftest-traj-gemini-analysis
Viewer • Updated • 1.76k • 57
violetxi/clbench-poker-halftest-traj-qwen3-4b-analysis
Viewer • Updated • 55 • 56
violetxi/clbench-poker-halftest-traj-qwen3-4b-nowm
Viewer • Updated • 110 • 23
violetxi/clbench-poker-halftest-traj-gemini-nowm
Viewer • Updated • 495 • 35
violetxi/clbench-exploitable-poker-wm-sft-future-summary
Viewer • Updated • 956 • 42
violetxi/clbench-exploitable-poker-wm-sft-next-state
Viewer • Updated • 1.04k • 40