messages listlengths 9 95 | metadata dict |
|---|---|
[
{
"role": "system",
"content": "You are a helpful assistant that can interact with a computer.\n\nYour response must contain exactly ONE bash code block with ONE command (or commands connected with && or ||).\nInclude a THOUGHT section before your command where you explain your reasoning process.\nFormat yo... | {
"id": "agent_task_20260715T162724056782Z",
"project": "assimp",
"vulnerability_id": "vulnerability_2",
"task_mode": "patch"
} |
[
{
"role": "system",
"content": "You are a helpful assistant that can interact with a computer.\n\nYour response must contain exactly ONE bash code block with ONE command (or commands connected with && or ||).\nInclude a THOUGHT section before your command where you explain your reasoning process.\nFormat yo... | {
"id": "agent_task_20260715T162724064053Z",
"project": "assimp",
"vulnerability_id": "vulnerability_4",
"task_mode": "patch"
} |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T162724055158Z","project":"assimp","vulnerability_id":"vulnerability_0","t(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T162724064214Z","project":"assimp","vulnerability_id":"vulnerability_3","t(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T162724064945Z","project":"assimp","vulnerability_id":"vulnerability_v643_(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T162724056931Z","project":"assimp","vulnerability_id":"vulnerability_1","t(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T171451262823Z","project":"bluez","vulnerability_id":"vulnerability_1","ta(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T171451264923Z","project":"bluez","vulnerability_id":"vulnerability_0","ta(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T171451251293Z","project":"bluez","vulnerability_id":"vulnerability_v643_2(...TRUNCATED) |
[{"role":"system","content":"You are a helpful assistant that can interact with a computer.\n\nYour (...TRUNCATED) | {"id":"agent_task_20260715T171451260245Z","project":"bluez","vulnerability_id":"vulnerability_v643_1(...TRUNCATED) |
End of preview. Expand in Data Studio
CyberForge Teacher Trajectories (GPT-5.4-mini)
1276 agentic security-patch trajectories from the GPT-5.4-mini teacher, cleansed to the final versions used to train the student models in the CyberForge paper.
Each line is one trajectory (JSONL): messages (system / user / assistant turns of the
mini-swe-agent loop) and metadata.
- Teacher: GPT-5.4-mini teacher
- Records: 1276
- Format: JSONL, one trajectory per line
Related
- Companion teacher set: cyberforge-teacher-traj-gemma4-31b
- Framework: CyberForge
- Training code: CyberForge-training
- Vulnerability dataset: CyberForge-projects
- Downloads last month
- 42