Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot extract the features (columns) for the split 'train' of the config 'default' of the dataset.
Error code:   FeaturesError
Exception:    ArrowInvalid
Message:      JSON parse error: The document is empty.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 280, in _generate_tables
                  df = pandas_read_json(f)
                       ^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 34, in pandas_read_json
                  return pd.read_json(path_or_buf, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 791, in read_json
                  json_reader = JsonReader(
                                ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 905, in __init__
                  self.data = self._preprocess_data(data)
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/pandas/io/json/_json.py", line 917, in _preprocess_data
                  data = data.read()
                         ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/utils/file_utils.py", line 844, in read_with_retries
                  out = read(*args, **kwargs)
                        ^^^^^^^^^^^^^^^^^^^^^
                File "<frozen codecs>", line 322, in decode
              UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 42: invalid start byte
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 247, in compute_first_rows_from_streaming_response
                  iterable_dataset = iterable_dataset._resolve_features()
                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 4196, in _resolve_features
                  features = _infer_features_from_batch(self.with_format(None)._head())
                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2533, in _head
                  return next(iter(self.iter(batch_size=n)))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2711, in iter
                  for key, pa_table in ex_iterable.iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2249, in _iter_arrow
                  yield from self.ex_iterable._iter_arrow()
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 494, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 384, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 283, in _generate_tables
                  raise e
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/json/json.py", line 246, in _generate_tables
                  pa_table = paj.read_json(
                             ^^^^^^^^^^^^^^
                File "pyarrow/_json.pyx", line 342, 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
              pyarrow.lib.ArrowInvalid: JSON parse error: The document is empty.

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.

M³Eval: Multi-Modal Memory Evaluation through Cognitively-Grounded Video Tasks

Paper Project Page Hugging Face Dataset GitHub Code

Jie Huang1,*Ruixun Liu1,*Sirui Sun1Xinyi Yang1Yin Li2Yixin Zhu1Yiwu Zhong1,†

1Peking University  2University of Wisconsin-Madison

* Equal contribution. † Corresponding author.

News

  • 2026-6-4: We released the M³Eval benchmark, code, and project page.

M³Eval Overview

M³Eval overview

Abstract

As multi-modal models advance towards long-form video understanding, memory emerges as a critical capability. Despite substantial effort in developing video datasets and benchmarks, existing work primarily focuses on perception and reasoning, without systematically evaluating memory: what models retain, how faithfully information is preserved, and how robust memory remains under interference.

To address this gap, we introduce M³Eval, the first comprehensive evaluation framework and benchmark for probing different memory dimensions in multi-modal models. Grounded in cognitive psychology, our design features carefully constructed tasks isolating key aspects of memory. Leveraging M³Eval, we conduct extensive experiments across representative multi-modal models, revealing consistent weaknesses and distinctive behaviors.

We find that models struggle to maintain disentangled representations when processing parallel video streams, exhibit interference patterns differing substantially from those observed in human memory, ground memory sources more reliably in the spatial domain than the temporal domain, and demonstrate limited symbolic memory. Collectively, our benchmark provides a valuable resource for future research, whereas our findings highlight memory as a fundamental yet underexplored capability and offer insights for designing more effective memory mechanisms in multi-modal models.

Main Results

Divided Attention

Divided Attention main result

Accuracy (%) on three divided attention metrics under the split-screen setting without swaps and with frequent left/right swaps.

Memory Interference

Memory Interference main result

Proactive: first video (V1) interferes with recall of the second video (V2); retroactive: second video (V2) interferes with recall of the first video (V1). Delta denotes proactive minus retroactive.

Interleaved Events

Interleaved Events main result

Accuracy (%) on four interleaved reconstruction metrics.

N-Back

N-Back main result

Average accuracy under two symbolic attributes, scene and action, averaged over all K and N configurations.

Usage

Install

git clone https://github.com/PKU-VaLuE-Lab/m3eval.git
cd m3eval/lmms-eval
uv pip install -e ".[all]"
cd ..

Download the Dataset

The benchmark is composed from existing public benchmarks. Users must follow the original licenses and usage terms of each source dataset.

Download the dataset and unpack it into data/m3eval/:

huggingface-cli download PKU-VaLuE-Lab/m3eval \
  --repo-type dataset \
  --local-dir data/m3eval

bash data/m3eval/unpack_archives.sh

After unpacking, data/m3eval/ should contain:

data/m3eval/
├── qa_root/
├── questions/
├── nback/
└── videos/
    ├── interleaved/
    ├── memory_interference/
    ├── split_screen/
    └── nback/

Evaluation

The main public entry point is:

bash lmms-eval/scripts/run_m3eval_vllm.sh \
  --model_path /path/to/your/model \
  --task m3eval \
  --gpus 0 \
  --batch_size 1

For multi-GPU data-parallel evaluation, use:

bash lmms-eval/scripts/run_m3eval_sharded_vllm.sh \
  --model_path /path/to/your/model \
  --task m3eval \
  --gpus 0,1,2,3 \
  --num_processes 4 \
  --batch_size 1

For a quick smoke run:

bash lmms-eval/scripts/run_m3eval_vllm.sh \
  --model_path /path/to/your/model \
  --task m3eval_memory_interference \
  --gpus 0 \
  --limit 1 \
  --max_frame_num 8

Useful task names:

  • m3eval
  • m3eval_memory_interference
  • m3eval_split_screen
  • m3eval_interleaved
  • m3eval_nback

The script writes outputs to lmms-eval/output/m3eval/ by default. See lmms-eval/README_M3EVAL.md for the dataset layout and task names.

To convert lmms-eval outputs into paper-facing tables:

python lmms-eval/scripts/aggregate_m3eval_results.py \
  lmms-eval/output/m3eval/path/to/*_results.json

Dataset Examples

Divided Attention

Divided Attention dataset example

Simultaneous memory for two side-by-side videos.

Click to expand more examples

Memory Interference

Memory Interference dataset example

Interference between sequentially presented videos.

Interleaved Events

Interleaved Events dataset example

Memory reconstruction from temporally interleaved clips.

N-Back

N-Back dataset example

Decide whether the final clip matches the clip N positions earlier.

Citation

If you use M³Eval in your work, please cite:

@article{huang2026m3eval,
  title   = {M3Eval: Multi-Modal Memory Evaluation through Cognitively-Grounded Video Tasks},
  author  = {Huang, Jie and Liu, Ruixun and Sun, Sirui and Yang, Xinyi and Li, Yin and Zhu, Yixin and Zhong, Yiwu},
  journal = {arXiv preprint arXiv:2606.05008},
  year    = {2026}
}
Downloads last month
785

Paper for PKU-VaLuE-Lab/m3eval