Dataset Viewer
Auto-converted to Parquet Duplicate
conversations
listlengths
2
69
agent
stringclasses
1 value
model
stringclasses
1 value
model_provider
stringclasses
1 value
date
stringlengths
27
27
task
stringlengths
23
23
episode
stringclasses
17 values
run_id
stringclasses
1 value
trial_name
stringlengths
32
32
result
stringclasses
3 values
instruction
stringlengths
4
7.82k
verifier_output
stringlengths
613
2.02k
prompt_token_ids
listlengths
0
33
completion_token_ids
listlengths
0
33
logprobs
listlengths
0
33
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to search the codebase to(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.826008Z
crosscodeeval-java-2138
episode-7
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2138__Kf9CgUX
0.0
"Based on the code snippet provided, I can see this is a Java file that needs to be completed. The c(...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)
[[40,1144,310,2624,279,1970,3051,310,3418,279,5265,6632,321,5245,1429,303,411,31084,7693,1970,11,748(...TRUNCATED)
[[-0.20377779006958008,-0.0014993611257523298,-0.00007128449942683801,-0.7951661944389343,-1.0559530(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to understand the project(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.793857Z
crosscodeeval-java-2137
episode-2
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2137__92BXx8U
0.0
"Looking at the code context, I need to complete the `register` method in a Java file. Based on the (...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)
[[40,1144,310,3418,279,2313,5759,321,1423,5265,3425,310,4434,279,1970,9971,13,198,248069,271,248058,(...TRUNCATED)
[[-0.5968888998031616,-0.010389531962573528,-0.00042667769594118,-1.6713695526123047,-0.052685480564(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to explore the codebase t(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:12.053833Z
crosscodeeval-java-2136
episode-6
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2136__BU9XGpW
0.0
"Based on the code snippet provided, I can see this is a Java file that needs completion. The `boots(...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)
[[40,1144,310,12804,279,1970,3051,310,3418,279,5265,6632,321,5245,1518,25720,279,1970,13,198,248069,(...TRUNCATED)
[[-0.04438904672861099,-0.0006665143300779164,-0.00018392273341305554,-1.430586576461792,-0.01749219(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to explore the codebase t(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.835971Z
crosscodeeval-java-2135
episode-5
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2135__WiMvLpF
0.0
"Based on the code snippet provided, I can see this is a Java file that needs completion. The code i(...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)
[[40,1144,310,12804,279,1970,3051,310,3418,279,21753,6065,3622,11,5261,2009,523,11,321,5194,523,310,(...TRUNCATED)
[[-0.11280428618192673,-0.00646025687456131,-0.00007939006172819063,-0.7626373171806335,-0.001975608(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to read the solution file(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.975088Z
crosscodeeval-java-2134
episode-5
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2134__UM5nGoL
0.0
Let me explore the project structure to understand the related files and complete the code.
"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)
[[40,1144,310,1301,279,6093,999,310,1436,279,1428,1528,321,3418,1092,3749,310,381,7888,13,198,248069(...TRUNCATED)
[[-0.42336589097976685,-0.01310333888977766,-0.00019596057245507836,-2.1547622680664062,-0.044191632(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>The user wants me to complete Ja(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:12.179521Z
crosscodeeval-java-2133
episode-1
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2133__BmBWk6o
0.0
"I notice there's a mismatch - the code you've provided is Java (package `com.wentong.jimu...`), but(...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)
[[760,1156,6587,728,310,4434,7693,1970,11,694,781,2908,4439,728,310,3462,424,310,264,641,3181,999,47(...TRUNCATED)
[[-0.7022656202316284,-0.11058884859085083,-0.3823295533657074,-0.007718621287494898,-5.960430826235(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>The user wants me to complete a (...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:12.122521Z
crosscodeeval-java-2132
episode-3
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2132__BZbwB5B
0.0
"Based on the code context provided, I can see this is a Java `ServiceThread` class that uses a `Cou(...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)
[[760,1156,6587,728,310,4434,264,7693,999,318,1856,12654,8,440,5144,13797,2193,13,6558,728,1118,2624(...TRUNCATED)
[[-0.7738001942634583,-0.03742692247033119,-0.6775264739990234,-0.015228915959596634,-0.000021338186(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to understand the project(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.947072Z
crosscodeeval-java-2131
episode-5
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2131__ffAAMLj
0.0
"Based on the code structure and the incomplete method `validateOrgType()`, I can analyze what needs(...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)
[[40,1144,310,3418,279,2313,5759,321,1423,279,22222,6632,310,4434,279,9307,41027,900,1654,13,198,248(...TRUNCATED)
[[-0.022877579554915428,-0.00007998607907211408,-8.225389137805905e-6,-1.2498832941055298,-0.0057460(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to read the context file (...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:11.818696Z
crosscodeeval-java-2130
episode-5
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2130__5wWbi3k
0.0
"Looking at the incomplete code, I can see it's a Java file (not Python) that needs to be completed.(...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)
[[40,1144,310,1301,279,2193,999,310,3418,279,1970,5759,321,1179,4434,279,31084,1654,539,2614,279,614(...TRUNCATED)
[[-0.48144710063934326,-0.004218251910060644,-0.00005209300070418976,-0.5698874592781067,-0.04692798(...TRUNCATED)
[{"role":"user","content":""},{"role":"assistant","content":"<think>I need to search the codebase to(...TRUNCATED)
opencode
vllm/7575007329916598
vllm
2026-08-07T19:31:12.268846Z
crosscodeeval-java-2129
episode-5
0271c327-4b88-4f10-869a-8f8b2d901303
crosscodeeval-java-2129__Tu772HV
0.0
"Based on the code snippet provided, I can see this is a Java file (OrgBuilder.java), not Python. Th(...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)
[[40,1144,310,2624,279,1970,3051,310,3418,279,32589,523,5759,11,279,22222,23344,11,321,1204,4283,464(...TRUNCATED)
[[-0.11573542654514313,-0.0008257782901637256,-0.00001943087590916548,-2.002714157104492,-0.31430247(...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: 7575007329916598

See tokenizer_provenance.json for a machine-readable version.

Downloads last month
34