Failing to Explore: Language Models on Interactive Tasks
Paper
•
2601.22345
•
Published
•
2
task_type
string | instance_index
int64 | config
string |
|---|---|---|
hill
| 0
|
{"max_value": 20, "function_type": "gauss_hills", "function_args": {"mu": [1.2835944367436807, 2.646676033502114, 3.831688772030403, 5.105581764683584, 6.437514316790188, 7.604339224488991, 8.771638676577446, 6.2], "var": [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.01], "mul": [1, 2, 4, 1, 2, 3, 4, 20]}}
|
hill
| 1
|
{"max_value": 20, "function_type": "gauss_hills", "function_args": {"mu": [1.3291117035923996, 2.7726013258516065, 4.012202040034237, 5.318805122062997, 6.450747721795136, 7.824902554509671, 8.948450001450865, 1.3], "var": [0.1, 0.2, 0.1, 0.1, 0.1, 0.2, 0.1, 0.01], "mul": [1, 5, 2, 1, 2, 3, 4, 20]}}
|
Explore-Exploit Bench — Failing to Explore: Language Models on Interactive Tasks
We refer to our github repo for more information on instances and how to use the dataset.
- Downloads last month
- 18
Size of downloaded dataset files:
499 kB
Size of the auto-converted Parquet files:
499 kB
Number of rows:
180