Adapter dataset cards
The source registry and coverage inventory are authoritative for revisions and counts. These short cards explain semantic mapping choices.
Software-engineering conversations
Adapters: swegym_openhands, swegym_sampled,
nvidia_open_swe_traces, swebench_verified_trajectories,
swesmith_trajectories.
System/user/assistant/tool messages and native instance IDs are preserved.
String-encoded NVIDIA messages/tools/metadata are JSON-decoded losslessly.
mini-SWE-agent call IDs are namespaced only when the source reuses an ID; the
original value remains in metadata. A final unresolved submit call may be a
valid terminal action when the source publishes its execution verdict, but no
tool result is fabricated. Gold patches remain raw provenance and are excluded
from SFT model_input.
Step-shaped SWE and terminal traces
Adapters: swebench_multilingual_trajectories,
terminalbench_trajectories, terminaltraj.
Every source action becomes a supervised assistant event and every published
response/state/terminal output becomes its causal environment event.
Terminal-Bench trials with steps=null are rejected before canonical import;
short or inadequately contextualized traces remain tier-5 diagnostics. Source
reward, cost, duration, model and scaffold remain independent metadata.
Tool and user-simulation environments
Adapters: agentsuite_tau2, official_tau3, toolbench.
Calls and results remain structured and linked. AgentSuite τ² correctness uses
the published evaluator fields. Official τ³ trials preserve reward evidence,
seed and trial and are protected evaluation data. ToolBench uses only the last
cumulative train_messages history so prefixes are not duplicated; its
win/valid_data flags are not treated as independent semantic proof.
agentsuite_bfcl is a diagnostic adapter: it proves the 115,466-row release
contains final function-call answers and evaluator scores but no returned tool
observations. These records fail the trajectory gate and are excluded from the
final library.
Web interaction
Adapters: mind2web, browsecomp_rollouts.
Mind2Web pairs each human operation with its full pre-action DOM; it has no binary task verdict. BrowseComp preserves research-agent search calls, full returned content and the source answer-judge verdict. Reference answers and judge explanations stay outside model input. BrowseComp rollouts are protected evaluation data.
ALFWorld
The adapter preserves task/scene data, human language annotations, PDDL high-level actions, low actions and image references. The audited release does not publish action-paired observations in the planner record, so mapped samples are explicitly incomplete and are not included in the qualifying library.