datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
rl__24GPU_base__mix_h2_language_balanced__r2egym-nl2bash-stackrl__24GPU_base__swe_rebench_patched_oracle__r2egym-nl2bash-stackDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k_glof5f5a963DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-5ep617edde0rl__24GPU_base__code-contests-noblock__r2egym-nl2bash-stackDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k_glob0c37db0DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k_glocbb0258aDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-5ep7724f234DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-5ep2cc93c10DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-5epcd171a23DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-8ep5f9af91aDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-7ep59daa7a4DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-wr01775995aDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-wr03a52d518DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-wr09d83cf2aDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4_6-traces-32ep-32k-wr00d517088rl__24GPU_shaped__swe_rebench_patched_oracle__r2egym-nl2bash-stackDCAgent2_terminal_bench_2_DCAgent_nl2bash-GLM-4.6-traces_Qwen3-8B_20251120_153822DCAgent2_terminal_bench_2_penfever_nl2bash_verified_gpt-5-nano-traces-restore-hbe5c1c2cdcagent2-terminal-bench-2-dcagent-nl2bash-glm-4-6-traces-qwen3-8b-20251118-115226nl2bashnl2bash-tasks-cleaned-oracle-qwen3.5-122b-131k-opencode-traces
Agent trace dataset
Decoding the literal token IDs
The prompt_token_ids / completion_token_ids / logprobs columns are the
verbatim tokens the serving engine emitted, stored PER AGENT STEP as a
list-of-lists (one inner list per turn). To turn them back into text you MUST
use the exact tokenizer the model was served with — a generic same-family
tokenizer will decode word tokens to garbage.
Served model / tokenizer source: Qwen/Qwen3.5-122B-A10B-FP8
from transformers… See the full description on the dataset page: https://huggingface.co/datasets/open-athena/nl2bash-tasks-cleaned-oracle-qwen3.5-122b-131k-opencode-traces.glm52-datagen-r11-07-nl2bash-tracesDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k-ab1674537b6DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k-ab13d395731terminal_bench_2_a3_rl_DCAgent2_nl2bash_tasks_cleaned_oracle_40_8B_20260830_081303DCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k-ab18e4c9bdfDCAgent2_terminal_bench_2_DCAgent2_nl2bash-verified-GLM-4.6-traces-32ep-32k-ab116928ab6rl__24GPU_base__mix_h2_language_proportional__r2egym-nl2bash-stackDCAgent2_terminal_bench_2_DCAgent_nl2bash-Qwen3-Coder-480B-traces_Qwen3-8B_20259e54c645
