Dataset Viewer
Auto-converted to Parquet Duplicate
conversations
list
agent
string
model
string
model_provider
string
date
string
task
string
episode
string
run_id
string
trial_name
string
result
string
instruction
string
verifier_output
string
prompt_token_ids
list
completion_token_ids
list
logprobs
list
[{"content":"","role":"user"},{"content":"<think>I need to implement a minimal Python StyleTable met(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-10T16:42:34.609866Z
pr-2758
episode-22
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-2758__cWbMgS9
0.0
"I'll implement a minimal Python StyleTable metadata reader with a CLI. Let me start by exploring th(...TRUNCATED)
"Running tests...\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)
[[40,1144,310,4072,264,16866,12654,11574,2473,10836,6396,440,37978,11,748,1042,728,1118,1716,279,142(...TRUNCATED)
[[-0.3022508919239044,-0.07982707023620605,-0.000027656107704387978,-0.1364264041185379,-0.005771162(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to implement a CLI workflow for c(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-12T19:42:01.105743Z
pr-0721
episode-5
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-0721__j4ygWGU
AgentTimeoutError
"I'll implement the clustering CLI workflow. Let me start by exploring the current state of the `/ap(...TRUNCATED)
null
[[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED)
[[760,1156,6587,728,310,4072,264,37978,27370,364,56391,22958,3018,383,264,5821,5979,13,6558,728,1387(...TRUNCATED)
[[-0.5924429893493652,-0.0059484574012458324,-0.002880827523767948,-0.00016377987049054354,-1.192090(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>I need to implement a Python CLI for validating and(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-10T13:36:42.317672Z
pr-2918
episode-4
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-2918__dgU3ggc
AgentTimeoutError
"I'll implement a Python CLI for validating and normalizing YAML membership specs. Let me start by e(...TRUNCATED)
null
[[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED)
[[40,1144,310,4072,264,12654,37978,364,66350,321,4472,4694,51342,15403,31184,471,1042,728,1118,12804(...TRUNCATED)
[[-1.0368025302886963,-0.08524812757968903,-0.0006498362636193633,-1.0379945039749146,-0.01934531331(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to create a self-contained Python(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-10T02:47:08.629708Z
pr-3614
episode-0
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-3614__BkN3sjU
AgentTimeoutError
"I'll implement this velocity controller benchmark. Let me first check the current state of the /app(...TRUNCATED)
null
[[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED)
[[760,1156,6587,728,310,1791,264,638,91938,12654,27502,5148,364,264,14984,6259,440,12804,5972,3627,1(...TRUNCATED)
[[-0.4332628548145294,-0.008639576844871044,-0.0035997102968394756,-0.00046087554073892534,-7.152541(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>This is a complex task to implement a staging-to-sa(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-09T15:44:59.632049Z
pr-4041
episode-6
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-4041__hEBCuCw
AgentTimeoutError
"I'll implement this staging-to-samples JSON anonymization pipeline. Let me start by exploring the c(...TRUNCATED)
null
[[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED)
[[1919,369,264,6150,3274,310,4072,264,46170,4534,1284,3888,4566,30128,1954,14855,13,6558,728,1387,14(...TRUNCATED)
[[-0.6731480360031128,-0.007646932266652584,-0.00014172960072755814,-1.1006354093551636,-0.256756812(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to create a Node.js CLI tool for (...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-12T03:23:45.700292Z
pr-1631
episode-40
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-1631__DxABmN7
0.0
"Implementation complete. Created a self-contained Node.js CLI for generating two-page PDF attestati(...TRUNCATED)
"Running tests...\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)
[[760,1156,6587,728,310,1791,264,5829,2764,37978,5224,364,22440,11030,83917,778,13,6558,728,1151,539(...TRUNCATED)
[[-0.4506203830242157,-0.007759197149425745,-0.0033061604481190443,-0.0001239697594428435,-5.9604508(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>I need to implement a Python CLI tool for collectin(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-11T20:31:50.725770Z
pr-2093
episode-0
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-2093__hVXSaas
AgentTimeoutError
null
null
[[248045,8678,198,2,13455,271,2523,599,2528,310,279,2614,5568,25,271,27,15449,29,198,4754,1267,763,3(...TRUNCATED)
[[40,1144,310,4072,264,12654,37978,5224,364,25068,3108,15796,471,1042,728,1151,539,6611,264,11466,11(...TRUNCATED)
[[-0.36979836225509644,-0.115858294069767,-0.00003957661101594567,-0.23904383182525635,-0.3313413560(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to create a Rust CLI tool that mo(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-12T10:55:10.848605Z
pr-1189
episode-66
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-1189__CtMSa4P
0.0
"I'll create a Rust CLI tool that models and prints a simplified ICMS GrupoIcms structure. Let me se(...TRUNCATED)
"Running tests...\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)
[[760,1156,6587,728,310,1791,264,32671,37978,5224,421,3983,321,22741,264,42349,18705,4671,72355,40,2(...TRUNCATED)
[[-0.23618823289871216,-0.030202750116586685,-0.015069345943629742,-0.0008199415751732886,-7.1525414(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to implement a self-contained SVG(...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-09T11:50:39.667166Z
pr-4321
episode-18
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-4321__A7SFK5q
1.0
"The SVG optimizer is implemented and working. Here's what was created:\n\n**Structure:**\n```\n/app(...TRUNCATED)
"Running tests...\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)
[[760,1156,6587,728,310,4072,264,638,91938,38479,24816,303,5829,2764,440,264,8734,5792,17120,13,6558(...TRUNCATED)
[[-1.7780046463012695,-0.0638117715716362,-0.0054302336648106575,-0.01591293141245842,-1.54971621668(...TRUNCATED)
[{"content":"","role":"user"},{"content":"<think>The user wants me to implement a local static site (...TRUNCATED)
opencode
vllm/1354961408891683
vllm
2026-07-12T06:58:07.065433Z
pr-1394
episode-12
06ca39e5-eacb-4dc4-8956-58c6f8139733
pr-1394__u5mFULL
1.0
"I'll implement this local static site deployment exercise. Let me start by exploring the current st(...TRUNCATED)
"Running tests...\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)
[[760,1156,6587,728,310,4072,264,2136,1064,2660,22449,9856,13,6558,728,1387,1441,1092,3749,310,381,2(...TRUNCATED)
[[-0.491271048784256,-0.006724704522639513,-0.005498403683304787,-9.53667131398106e-6,-3.57627072844(...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: 1354961408891683

See tokenizer_provenance.json for a machine-readable version.

Downloads last month
42