Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,37 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+eval_results/global_step_0/mmlu_stem/test_qwen-boxed_-1_seed0_t1.0_s0_e-1.jsonl filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_acc_pass_acc.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_acc_keywords.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_acc_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_avg_stop_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_box_ratio_and_token_length.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_clip_ratio.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_repeat_ratio_and_token_length.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_correct_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_tokens_keywords.png filter=lfs diff=lfs merge=lfs -text
+eval_results/plots/eval_results_wrong_tokens.png filter=lfs diff=lfs merge=lfs -text
+global_step_40/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_50/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_60/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_70/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_80/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_90/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_acc_keywords.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_avg_stop_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_acc_pass_acc.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_box_ratio_and_token_length.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_acc_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_correct_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_tokens_keywords.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_repeat_ratio_and_token_length.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_wrong_tokens.png filter=lfs diff=lfs merge=lfs -text
+eval_results_avg4/plots/eval_results_avg4_clip_ratio.png filter=lfs diff=lfs merge=lfs -text
+eval_results_ood/global_step_0/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_0_actor_huggingface/2025-08-19T03-03-31.558565/details_extended|ifeval|0_2025-08-19T03-03-31.558565.csv filter=lfs diff=lfs merge=lfs -text
+eval_results_ood/global_step_10/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_10_actor_huggingface/2025-08-19T02-59-44.015414/details_extended|ifeval|0_2025-08-19T02-59-44.015414.csv filter=lfs diff=lfs merge=lfs -text
+eval_results_ood/global_step_20/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_20_actor_huggingface/2025-08-19T02-45-28.477295/details_extended|ifeval|0_2025-08-19T02-45-28.477295.csv filter=lfs diff=lfs merge=lfs -text
+global_step_10/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_100/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_20/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+global_step_30/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text

eval_results/global_step_0/mmlu_stem/test_qwen-boxed_-1_seed0_t1.0_s0_e-1.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c128f44b9bd5065604edd55ea62d32b25dfd015eac122620303e55215bed515
+size 12239207

eval_results_ood/global_step_0/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_0_actor_huggingface/2025-08-19T03-03-31.558565/details_extended|ifeval|0_2025-08-19T03-03-31.558565.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77f4cbd1d57baaa65840d41c4e3fdf408aa0a5d078c3ab233d451de97e99b727
+size 18592855

eval_results_ood/global_step_0/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_0_actor_huggingface/2025-08-19T03-03-31.558565/details_extended|ifeval|0_2025-08-19T03-03-31.558565.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d611194575c562264b41400d3d9288f43e8c6320cb0e6d912a6489b1d85682f1
+size 2789061

eval_results_ood/global_step_0/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_0_actor_huggingface/2025-08-19T03-03-31.558565/details_lighteval|gpqa:diamond|0_2025-08-19T03-03-31.558565.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4073c119302a1e9cd694f357f710227a8957303f7178eb8702a4bd2c42b5a7b
+size 566842

eval_results_ood/global_step_10/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_10_actor_huggingface/2025-08-19T02-59-44.015414/details_extended|ifeval|0_2025-08-19T02-59-44.015414.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50f4fa9aca026027eac08f0ed0ddc2b00d926e18b7f88e330fc4293c392a73d0
+size 17748213

eval_results_ood/global_step_10/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_10_actor_huggingface/2025-08-19T02-59-44.015414/details_extended|ifeval|0_2025-08-19T02-59-44.015414.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cea5449f92cc18f780f33cf1ab76f3769bac4d58eb2e93547daf41df8842abeb
+size 3026523

eval_results_ood/global_step_10/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_10_actor_huggingface/2025-08-19T02-59-44.015414/details_lighteval|gpqa:diamond|0_2025-08-19T02-59-44.015414.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f310b31e7299ff4348cd8d9655f8334f9b76a8990f85a17be0445fc50aafe21c
+size 611170

eval_results_ood/global_step_100/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_100_actor_huggingface/2025-08-19T03-23-25.775217/details_extended|ifeval|0_2025-08-19T03-23-25.775217.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81c4fdf2936ca8d92b10ccc2b75ceb72f551b3b3be588ede7604967537743b29
+size 1526557

eval_results_ood/global_step_100/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_100_actor_huggingface/2025-08-19T03-23-25.775217/details_lighteval|gpqa:diamond|0_2025-08-19T03-23-25.775217.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a6c707d324d9afd6a1cfd752d177e675674484b761e2760fe7ce65b56324180
+size 762665

eval_results_ood/global_step_20/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_20_actor_huggingface/2025-08-19T02-45-28.477295/details_extended|ifeval|0_2025-08-19T02-45-28.477295.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0309fa76751880316815296f3f5d9ea697b4bcf47aab8ff480b5b00cb1728aa9
+size 17747706

eval_results_ood/global_step_20/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_20_actor_huggingface/2025-08-19T02-45-28.477295/details_extended|ifeval|0_2025-08-19T02-45-28.477295.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d101c8fa4c97ffd37e3554371ec26e987a62ad1b7d8cd2551e0cecae28d65d5
+size 2445648

eval_results_ood/global_step_20/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_20_actor_huggingface/2025-08-19T02-45-28.477295/details_lighteval|gpqa:diamond|0_2025-08-19T02-45-28.477295.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1dffd5da7c2e6e280733a688198e964c5b843c36ac5e1187658f3a1da0467b32
+size 604606

eval_results_ood/global_step_30/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_30_actor_huggingface/2025-08-19T02-31-43.587146/details_extended|ifeval|0_2025-08-19T02-31-43.587146.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e06532f2a4ca0d6fc447b2c01eaa81246f747d6e82cd3a415179fd94e01ab5c9
+size 1997084

eval_results_ood/global_step_30/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_30_actor_huggingface/2025-08-19T02-31-43.587146/details_lighteval|gpqa:diamond|0_2025-08-19T02-31-43.587146.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49be0512c751e57b2bdc48244427efe32f3db3411839a30cc75027c2a0bff29a
+size 659079

eval_results_ood/global_step_40/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_40_actor_huggingface/2025-08-19T02-32-20.502116/details_extended|ifeval|0_2025-08-19T02-32-20.502116.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccbccd663da32f237c308c7ce146a10b1403cd60ade911b955a8466c59aa4315
+size 1933203

eval_results_ood/global_step_40/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_40_actor_huggingface/2025-08-19T02-32-20.502116/details_lighteval|gpqa:diamond|0_2025-08-19T02-32-20.502116.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72fe64ee7be245e4906a6f79dea95073c247311a70f7469a2e86d7ca3d721dc1
+size 729678

eval_results_ood/global_step_50/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_50_actor_huggingface/2025-08-19T02-27-37.276111/details_extended|ifeval|0_2025-08-19T02-27-37.276111.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7d7fc6fb913cabb9c3bb4ab6faf1d413a69f88cb882fbd63f23a0ce21d85ad8
+size 1887216

eval_results_ood/global_step_50/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_50_actor_huggingface/2025-08-19T02-27-37.276111/details_lighteval|gpqa:diamond|0_2025-08-19T02-27-37.276111.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4626ac4a1695080f6379b1712918e3c90c9c9b2d6bb41ce561ffc262554982e1
+size 710992

eval_results_ood/global_step_60/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_60_actor_huggingface/2025-08-19T02-24-48.573923/details_extended|ifeval|0_2025-08-19T02-24-48.573923.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35ce730b988707e09af26e33ee0b669dc0b0eac8a56c63d1446260dd37bc6d1c
+size 1633920

eval_results_ood/global_step_60/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_60_actor_huggingface/2025-08-19T02-24-48.573923/details_lighteval|gpqa:diamond|0_2025-08-19T02-24-48.573923.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b3eafdc4a5dc80faae8416428e649a5ea69b3720ab02eb6945c3455ccfd8379
+size 717748

eval_results_ood/global_step_70/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_70_actor_huggingface/2025-08-19T02-23-41.225302/details_extended|ifeval|0_2025-08-19T02-23-41.225302.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4ff941a9a8eb14c97cd6a36795beb4e5cddcf45ac9977f605eb78a892b75e1b
+size 1554610

eval_results_ood/global_step_70/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_70_actor_huggingface/2025-08-19T02-23-41.225302/details_lighteval|gpqa:diamond|0_2025-08-19T02-23-41.225302.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:896ae09b2e2167fe2443b2586ec718a9a92a7eb5a521b891fdf654b8b0a7f06a
+size 711115

eval_results_ood/global_step_80/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_80_actor_huggingface/2025-08-19T03-22-27.759640/details_extended|ifeval|0_2025-08-19T03-22-27.759640.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d7564a8979cd22c865885d2905e4d6999907884b7c4139c2e8582799de3f041
+size 1863422

eval_results_ood/global_step_80/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_80_actor_huggingface/2025-08-19T03-22-27.759640/details_lighteval|gpqa:diamond|0_2025-08-19T03-22-27.759640.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e3309afb1ba407363c2582463ed55d238c47b3f50af967b601c0a0bd54572b0
+size 731554

eval_results_ood/global_step_90/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_90_actor_huggingface/2025-08-19T03-23-41.379521/details_extended|ifeval|0_2025-08-19T03-23-41.379521.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bdc3b985034af8c1a96535639333cd8fec96b045a5cd0147f1ccc7f09931fab
+size 1712739

eval_results_ood/global_step_90/details/_home_work_compass_innovation_minzijun_checkpoints_verl_role_sft_grpo_deepmath_train_sample_6144_context_4k_Qwen3-1.7B-base_max_response4096_batch1024_rollout8_vllm_True_bias0.3_global_step_90_actor_huggingface/2025-08-19T03-23-41.379521/details_lighteval|gpqa:diamond|0_2025-08-19T03-23-41.379521.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a22384154eb28672a4c38af7f8b186cbaecdf854de6d2a9596b0bd3293aa4e5
+size 761415

global_step_0/actor/huggingface/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6df85b39330e5a425ee36253d0f894e4387e4f0a15b9c53cb467d668e6b3a841
+size 3441185608

global_step_10/actor/huggingface/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

global_step_10/data.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5d7cb574e77c186989ac15b0b93bfabe487df954f0d0061aeabf2d618cedd25
+size 1947