arxiv:2609.07064
Sohee Kim
joyhee
AI & ML interests
None yet
Recent Activity
upvoted a paper 5 days ago
Eat Your Precision: Precision Objective Mitigates Reward Hacking in Rubric-based RL upvoted a paper 17 days ago
ACCRUE: Continual Off-Policy Reinforcement Learning for Robot Foundation ModelsOrganizations
None yet
models 196
joyhee/Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s5_exp1-only-ep3-step27
4B • Updated • 7
joyhee/Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s5_exp1-only-ep3-step26
4B • Updated • 5
joyhee/again-Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s4_exp2-only-ep3-step27
4B • Updated • 5
joyhee/again-Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s4_exp2-only-ep3-step12
4B • Updated • 4
joyhee/Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s2_exp-only-ep3-step27
4B • Updated • 5
joyhee/Qwen2.5-VL-3B-Instruct-GRPO-ver3-v6_s2_exp-only-ep3-step24
4B • Updated • 5
joyhee/Qwen2.5-VL-7B-Instruct-GRPO-ver3-v6_s2_exp_s4_exp2_s5_exp1_only_blocks_ver3-len-min_100-step29
8B • Updated • 9
joyhee/Qwen2.5-VL-7B-Instruct-GRPO-ver3-v6_s2_exp_s4_exp2_s5_exp1_only_blocks_ver3-len-min_100-s28
8B • Updated • 10
joyhee/again-Qwen2.5-VL-7B-Instruct-GRPO-ver3-v6_s2_exp_s4_exp2_s5_exp1_only_blocks_ver3-step29
8B • Updated • 7
joyhee/again-Qwen2.5-VL-7B-Instruct-GRPO-ver3-v6_s2_exp_s4_exp2_s5_exp1_only_blocks_ver3-step26
8B • Updated • 7
datasets 8
joyhee/spbench-full
Viewer • Updated • 11.6k • 45
joyhee/rl_asset_v6_ver3_s245_nobg-only_blocks-full
Viewer • Updated • 52k • 8
joyhee/SpatialLadder-str
Viewer • Updated • 20.7k • 12
joyhee/rl_scene-str
Viewer • Updated • 7.98k • 20
joyhee/rl_full_v6_tot_nobg-only_blocks-str
Viewer • Updated • 164k • 26
joyhee/rl_S5_link_scene-str
Viewer • Updated • 4.19k • 7
joyhee/rl_mindcube-str
Viewer • Updated • 11.1k • 8
joyhee/rl_full_v6_tot_nobg-str
Viewer • Updated • 93.6k • 5