CoolFace
Datasetpublic

open-athena/exp_rpt_crosscodeeval-java-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/exp_rpt_crosscodeeval-java-qwen3.5-122b-131k-opencode-traces.

sourceHugging Faceupdated 3mo agoView on Hugging Face
0likes103downloads
../
filetrain-00000.parquet74.4 MBdownload
filetrain-00001.parquet67.9 MBdownload
filetrain-00002.parquet72.2 MBdownload
filetrain-00003.parquet65.3 MBdownload
filetrain-00004.parquet67.1 MBdownload
filetrain-00005.parquet65.8 MBdownload
filetrain-00006.parquet69.8 MBdownload
filetrain-00007.parquet66.0 MBdownload
filetrain-00008.parquet73.7 MBdownload
filetrain-00009.parquet66.6 MBdownload
filetrain-00010.parquet40.5 MBdownload

open-athena/exp_rpt_crosscodeeval-java-qwen3.5-122b-131k-opencode-traces · main · files are served by the source, never re-hosted here