Dataset Viewer
Auto-converted to Parquet Duplicate
policy
string
wm
string
task_name
string
task_category
string
task_difficulty
string
task_tags
list
task_instruction
string
num_turns
int64
messages
list
wm_prediction_correct
list
wm_pred_token_f1
list
test_names
list
test_passed
list
tests_passed
int64
tests_total
int64
solved
int64
history_rewritten
bool
summarysrc
none
bn-fit-modify
scientific-computing
hard
[ "bayesian-network", "stats" ]
Evaluates the ability to recover a Bayesian Network DAG structure from data, perform causal interventions, and sample from the modified network.
30
[ { "role": "user", "source": "env", "phase": "prompt", "content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell...
[ null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null ]
[ null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null, null ]
[ "test_outputs.py::test_bn_sample_exists", "test_outputs.py::test_learned_dag_structure_exists", "test_outputs.py::test_learned_dag_structure_csv_col_names", "test_outputs.py::test_learned_dag_structure", "test_outputs.py::test_intervened_dag_structure_exists", "test_outputs.py::test_intervened_dag_structu...
[ 1, 1, 1, 0, 1, 1, 0, 0, 0 ]
5
9
0
false
summarysrc
none
cancel-async-tasks
software-engineering
hard
[ "async", "concurrency", "python" ]
Evaluates the ability to implement async task concurrency control with proper cleanup on cancellation, including the edge case of queued tasks.
5
[ { "role": "user", "source": "env", "phase": "prompt", "content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell...
[ null, null, null, null, null ]
[ null, null, null, null, null ]
[ "test_outputs.py::test_run_py_file_exists", "test_outputs.py::test_tasks_run_concurrently", "test_outputs.py::test_tasks_obey_max_concurrent", "test_outputs.py::test_tasks_cancel_below_max_concurrent", "test_outputs.py::test_tasks_cancel_at_max_concurrent", "test_outputs.py::test_tasks_cancel_above_max_co...
[ 1, 1, 1, 1, 1, 0 ]
5
6
0
false
summarysrc
none
circuit-fibsqrt
software-engineering
hard
[ "software-engineering" ]
"Evaluates the agent's ability to implement complex mathematical functions (Fibonacci of integer squ(...TRUNCATED)
74
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
["test_outputs.py::test_gates_file_exists","test_outputs.py::test_gates_file_size","test_outputs.py:(...TRUNCATED)
[ 1, 1, 0 ]
2
3
0
true
summarysrc
none
cobol-modernization
software-engineering
easy
[ "coding" ]
"Evaluates the ability to reverse-engineer and reimplement a COBOL program's business logic in Pytho(...TRUNCATED)
30
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
["test_outputs.py::test_required_files_exist","test_outputs.py::test_data_files_exist","test_outputs(...TRUNCATED)
[ 1, 1, 0 ]
2
3
0
false
summarysrc
none
code-from-image
software-engineering
medium
[ "ocr" ]
"Evaluates an agent's ability to extract code from an image using OCR or vision models, implement th(...TRUNCATED)
83
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[ "test_outputs.py::test_output_file_exists", "test_outputs.py::test_output_is_correct" ]
[ 1, 0 ]
1
2
0
true
summarysrc
none
constraints-scheduling
personal-assistant
medium
[ "calendar", "scheduling", "constraint-satisfaction", "ics-parsing", "temporal-reasoning" ]
"Find an optimal 1-hour meeting slot for three people with complex availability constraints by parsi(...TRUNCATED)
68
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
["test_outputs.py::test_structure_event_title_attendees_duration_window","test_outputs.py::test_cons(...TRUNCATED)
[ 0, 0, 0 ]
0
3
0
true
summarysrc
none
custom-memory-heap-crash
debugging
medium
[ "cpp", "memory-management", "debugging" ]
"Evaluates the ability to debug and fix a C++ program that crashes in release mode due to a static i(...TRUNCATED)
30
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
["test_outputs.py::test_protected_files_not_modified","test_outputs.py::test_program_compiles_debug"(...TRUNCATED)
[ 1, 1, 1, 1, 0, 1 ]
5
6
0
false
summarysrc
none
db-wal-recovery
file-operations
medium
[ "database", "encryption", "recovery" ]
"Tests the ability to decrypt an XOR-encrypted SQLite WAL file and recover complete database content(...TRUNCATED)
30
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
["test_outputs.py::test_recovered_json_exists","test_outputs.py::test_recovered_json_is_valid","test(...TRUNCATED)
[ 0, 0, 0, 0, 0, 0, 0 ]
0
7
0
false
summarysrc
none
dna-assembly
scientific-computing
hard
[ "biology", "cloning" ]
"Evaluates the ability to design PCR primers for Golden Gate assembly by applying molecular biology (...TRUNCATED)
32
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[ "test_outputs.py::test_primers" ]
[ 0 ]
0
1
0
false
summarysrc
none
dna-insert
scientific-computing
medium
[ "biology", "cloning" ]
"Evaluates the ability to design PCR primers for site-directed mutagenesis by analyzing plasmid sequ(...TRUNCATED)
31
[{"role":"user","source":"env","phase":"prompt","content":"You are an AI assistant tasked with solvi(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null(...TRUNCATED)
[ "test_outputs.py::test_primers" ]
[ 0 ]
0
1
0
false
End of preview. Expand in Data Studio

No dataset card yet

Downloads last month
58