conversations listlengths 2 52 | agent stringclasses 1
value | model stringclasses 1
value | model_provider stringclasses 1
value | date stringlengths 27 27 | task stringlengths 23 23 | episode stringclasses 18
values | run_id stringclasses 1
value | trial_name stringlengths 32 32 | result stringclasses 3
values | instruction stringlengths 5 4.58k ⌀ | verifier_output stringlengths 613 2.02k ⌀ | prompt_token_ids listlengths 0 26 | completion_token_ids listlengths 0 26 | logprobs listlengths 0 26 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
[{"content":"","role":"user"},{"content":"<think>The user is asking me to complete a Java code snipp(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-09T11:33:31.747661Z | crosscodeeval-java-0055 | episode-4 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-0055__Z86a6Cp | 0.0 | "Looking at the code snippet, I need to complete the `getAnswer` method in the `BardClient` Java cla(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[760,1156,369,9859,728,310,4434,264,7693,1970,41643,364,264,61069,2865,523,11,694,781,1318,728,310,(...TRUNCATED) | [[-0.3897683024406433,-0.01787577196955681,-0.8471676707267761,-0.03500472754240036,-0.0033908744808(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>The user wants me to complete a Java file that's be(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-08T17:13:27.964487Z | crosscodeeval-java-1631 | episode-5 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-1631__bc9QevF | 0.0 | (tool use) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [] | [] | [] |
[{"content":"","role":"user"},{"content":"<think>I need to explore the project structure and underst(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-08T13:01:35.359402Z | crosscodeeval-java-1689 | episode-5 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-1689__E9BtyNz | 0.0 | "Looking at the incomplete Java code, I can see it's a GraphQL external field handler for a MockServ(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[40,1144,310,12804,279,2313,5759,321,3418,279,5265,3425,310,4434,411,7693,1970,9971,13,198,248069,2(...TRUNCATED) | [[-0.4832904636859894,-0.0009382124990224838,-0.00007593328336952254,-1.1899685859680176,-0.01310310(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>This is a Java code completion task, but the user w(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-08T13:04:18.619283Z | crosscodeeval-java-1673 | episode-1 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-1673__jG4Qviz | 0.0 | "Code completed. The missing logic was:\n- `if(album.getId() == music.getAlbum())` - to match the cu(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[1919,369,264,7693,1970,9463,3274,11,694,279,1156,6587,728,310,3462,279,6093,310,264,12654,999,471,(...TRUNCATED) | [[-1.059464454650879,-0.4167984426021576,-0.04754119738936424,-0.2238316386938095,-0.096064761281013(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>The user wants me to complete a Java code snippet, (...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-09T11:50:28.586263Z | crosscodeeval-java-0003 | episode-2 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-0003__vk8ApiC | 0.0 | "Looking at the code snippet, this is a Java file (not Python) that needs to complete a gRPC server (...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[760,1156,6587,728,310,4434,264,7693,1970,41643,11,694,781,2224,9859,728,310,1925,279,6093,303,264,(...TRUNCATED) | [[-0.43668681383132935,-0.12045224010944366,-0.10699179768562317,-0.004350956063717604,-9.4174693003(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>I need to understand the existing codebase patterns(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-09T11:42:46.341387Z | crosscodeeval-java-0027 | episode-6 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-0027__wd5LrRv | 0.0 | "Based on the code context provided, I can see this is a Java file for DingDing (钉钉) clock-in au(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[40,1144,310,3418,279,6149,1970,3051,12261,539,14603,364,4283,37066,321,279,11101,4070,12755,5245,1(...TRUNCATED) | [[-0.06302305310964584,-0.0008206565398722887,-0.000021338186343200505,-1.9925470352172852,-0.018545(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>I need to explore the project structure to understa(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-09T05:41:30.053541Z | crosscodeeval-java-0829 | episode-4 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-0829__egBAcbg | 1.0 | "Based on the code context provided, I can see this is a Netty-based message encoder for an RPC syst(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[40,1144,310,12804,279,2313,5759,310,3418,279,5265,6632,318,1835,2674,11,5952,3911,11,5952,5273,8,3(...TRUNCATED) | [[-0.26436007022857666,-0.0032914266921579838,-0.000024437558749923483,-1.9177693128585815,-0.015605(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>I need to read the existing solution file to unders(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-08T17:13:23.907100Z | crosscodeeval-java-1680 | episode-3 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-1680__cf93rPu | AgentTimeoutError | "I need to complete the Java code. Let me first search for related files to understand the context b(...TRUNCATED) | null | [] | [] | [] |
[{"content":"","role":"user"},{"content":"<think>I need to search for related files in the project t(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-09T06:57:51.925361Z | crosscodeeval-java-0478 | episode-4 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-0478__cR2tfsf | 0.0 | "Looking at the incomplete Java code, I can see the `saveFloat` method is cut off mid-write. The met(...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[40,1144,310,2624,364,5265,3425,303,279,2313,310,3418,279,2193,321,4434,279,1970,9971,13,198,248069(...TRUNCATED) | [[-0.24659311771392822,-0.014592600055038929,-0.00005268899258226156,-1.2236948013305664,-0.90643674(...TRUNCATED) |
[{"content":"","role":"user"},{"content":"<think>I need to search the codebase to understand the App(...TRUNCATED) | opencode | vllm/1500772956264735 | vllm | 2026-07-08T17:27:58.065283Z | crosscodeeval-java-1577 | episode-6 | d831ff5a-0f5f-46cc-847d-40957a77460a | crosscodeeval-java-1577__YoAVL2R | 0.0 | "Based on the code context provided, I can see this is a Java IntelliJ plugin settings configurable (...TRUNCATED) | "Running CrossCodeEval tests (python)...\n============================= test session starts ========(...TRUNCATED) | [[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED) | [[40,1144,310,2624,279,1970,3051,310,3418,279,1781,5896,2120,321,1781,5896,1349,6632,310,9971,4434,2(...TRUNCATED) | [[-0.94841468334198,-0.026369759812951088,-0.00010132257011719048,-1.044947624206543,-0.696038246154(...TRUNCATED) |
End of preview. Expand in Data Studio
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 import AutoTokenizer
# Use the served model's own tokenizer (pull from the ref above; if it is a
# gs:// mirror, copy tokenizer.json/tokenizer_config.json/vocab.json/merges.txt
# locally first and point AutoTokenizer at that dir).
tok = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-122B-A10B-FP8")
# token_ids are list-of-lists (one list per turn) — decode each turn:
text = [tok.decode(turn, skip_special_tokens=False) for turn in completion_token_ids]
Engine-reported served model name: 1500772956264735
See tokenizer_provenance.json for a machine-readable version.
- Downloads last month
- -