The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 291, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 306, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
post-deepswe-traj
SWE-bench Pro rollouts from mini-swe-agent 1.15.0, plus the pass/fail metadata for the same 731 instances.
The agent is mini-swe-agent (trajectory_format: mini-swe-agent-1), not the Codex harness and not SWE-agent. Each step is one THOUGHT plus one bash block, executed in the task Docker image with working directory /app and step_limit 250.
Files
| File | What it is |
|---|---|
swebench_pro_qwen3.8-27b_before_after_lora_and_deepseek-v4.1-flash_traj.tar.zst |
The three trajectory sets |
swebench_pro_qwen3.8-27b_before_after_lora_and_deepseek-v4.1-flash_score_metadata.json |
Which instances passed, failed, or were not scored |
Extract with:
tar -I zstd -xf swebench_pro_qwen3.8-27b_before_after_lora_and_deepseek-v4.1-flash_traj.tar.zst
Directories inside the archive:
qwen3.8-27b_before_lora— base Qwen3.8-27Bqwen3.8-27b_after_lora— the same base after LoRA on 203 DeepSeek-correct / Qwen-wrong trajectoriesdeepseek-v4.1-flash— DeepSeek-v4.1-flash
A trajectory is <instance_id>/<instance_id>.traj.json. That instance_id is the same key as in the metadata.
Scores
The denominator is submissions that contain diff --git . A scored instance passes only when every fail-to-pass and pass-to-pass test passed. not_scored still has a trajectory; it was left out of the denominator because the submission had no diff.
| Model | Passed | Failed | Not scored | Accuracy on scored |
|---|---|---|---|---|
| Qwen3.8-27B before LoRA | 432 | 232 | 67 | 432/664 = 65.06% |
| Qwen3.8-27B after LoRA | 558 | 112 | 61 | 558/670 = 83.28% |
| DeepSeek-v4.1-flash | 596 | 84 | 51 | 596/680 = 87.65% |
Nine before-LoRA trajectories live in qwen3.8-27b_before_lora/early_launch/. They are included in the 731 ids.
- Downloads last month
- 24