Datasets:

dnakhla
/

terminal-desk

Tasks:

Text Generation

Tags:

reinforcement-learning

Libraries:

License:

Dataset card Files Files and versions

2.03 MB

Ctrl+K

Ctrl+K

1 contributor

History: 9 commits

dnakhla's picture

v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks

1b83db2 verified 29 days ago

analysis
v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks 29 days ago
source
v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks 29 days ago
.gitattributes

2.59 kB
Terminal Desk v1: tasks, screen results, source about 1 month ago
README.md

18.9 kB
v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks 29 days ago
screen2_qwen3.8-27b.jsonl

133 kB
terminal-desk v2: tbench2_env interface mirror, artifact tasks, budget-emulated screen about 1 month ago
screen_v3.jsonl

157 kB
v3: 24 SkillsBench-shaped families, 288-episode screen, 50 selected ids, pre-registration about 1 month ago
screen_v3_fixed.jsonl

122 kB
v4: decisiveness gate (K=3), 17 office/software/finance tasks labelled by the arena's own run; v3c result; corrected screen about 1 month ago
selected_tasks.txt

706 Bytes
v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks 29 days ago
tasks.jsonl

138 kB
v4 result (3/40, office 2), v5 card: replicate under a new id, 21 tasks 29 days ago