source string | benchmark string | dataset_version_in_page_data list | task_count int64 | tasks list | github_repository string | github_link_note string |
|---|---|---|---|---|---|---|
https://www.tbench.ai/benchmarks/terminal-bench-1 | Terminal-Bench 1.0 | [
"0.1.1"
] | 80 | [
{
"task": "blind-maze-explorer-5x5",
"category": "games",
"difficulty": "hard",
"tags": [
"maze",
"exploration",
"mapping",
"algorithms"
],
"github_url": "https://github.com/harbor-framework/terminal-bench/tree/main/original-tasks/blind-maze-explorer-5x5",
"origin... | https://github.com/harbor-framework/terminal-bench | The page's laude-institute/tasks URLs are stale. Corrected URLs use the current harbor-framework/original-tasks paths, base-task paths for generated variants, or a fixed historical commit for removed tasks. |
https://www.tbench.ai/benchmarks/terminal-bench-2 | Terminal-Bench 2.0 | null | 89 | [
{
"task": "adaptive-rejection-sampler",
"category": "scientific-computing",
"difficulty": "medium",
"tags": [
"applied-statistics",
"adaptive-rejection-sampling",
"Bayesian-inference",
"simulation"
],
"github_url": null,
"original_page_github_url": null,
"gith... | null | null |
No dataset card yet
- Downloads last month
- 52