LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory
Paper • 2410.10813 • Published • 15
Datasets:
Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code: DatasetGenerationCastError
Exception: DatasetGenerationCastError
Message: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 9 new columns ({'answer_session_ids', 'answer', 'haystack_sessions', 'haystack_session_ids', 'question_date', 'haystack_dates', 'question_id', 'question_type', 'question'}) and 1 missing columns ({'sample_id'}).
This happened while the json dataset builder was generating data using
hf://datasets/justis-xu/memory-eval-zh/longmemeval-s/longmemeval_s_cleaned.json (at revision c4114ddcb789dad02ba59a757c7e08655c35317a), ['hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/pruned_questions.json'], ['hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/pruned_questions.json']
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1848, in _prepare_split_single
writer.write_table(table)
~~~~~~~~~~~~~~~~~~^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
self._write_table(pa_table, writer_batch_size=writer_batch_size)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
pa_table = table_cast(pa_table, self._schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
question_id: string
question_type: string
question: string
question_date: string
answer: extension<arrow.json>
answer_session_ids: list<item: string>
child 0, item: string
haystack_dates: list<item: string>
child 0, item: string
haystack_session_ids: list<item: string>
child 0, item: string
haystack_sessions: list<item: list<item: struct<role: string, content: string, has_answer: bool>>>
child 0, item: list<item: struct<role: string, content: string, has_answer: bool>>
child 0, item: struct<role: string, content: string, has_answer: bool>
child 0, role: string
child 1, content: string
child 2, has_answer: bool
event_summary: extension<arrow.json>
conversation: extension<arrow.json>
session_summary: extension<arrow.json>
observation: extension<arrow.json>
qa: null
-- schema metadata --
huggingface: '{"info": {"features": {"question_id": {"dtype": "string", "' + 899
to
{'qa': List(Json(decode=True)), 'conversation': Json(decode=True), 'event_summary': Json(decode=True), 'observation': Json(decode=True), 'session_summary': Json(decode=True), 'sample_id': Value('string')}
because column names don't match
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1850, in _prepare_split_single
raise DatasetGenerationCastError.from_cast_error(
...<4 lines>...
)
datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 9 new columns ({'answer_session_ids', 'answer', 'haystack_sessions', 'haystack_session_ids', 'question_date', 'haystack_dates', 'question_id', 'question_type', 'question'}) and 1 missing columns ({'sample_id'}).
This happened while the json dataset builder was generating data using
hf://datasets/justis-xu/memory-eval-zh/longmemeval-s/longmemeval_s_cleaned.json (at revision c4114ddcb789dad02ba59a757c7e08655c35317a), ['hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/pruned_questions.json'], ['hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/locomo/locomo10_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/longmemeval_s_cleaned_zh.json', 'hf://datasets/justis-xu/memory-eval-zh@c4114ddcb789dad02ba59a757c7e08655c35317a/longmemeval-s/pruned_questions.json']
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)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.
qa list | conversation string | event_summary string | observation string | session_summary string | sample_id string |
|---|---|---|---|---|---|
[
"{\"question\":\"When did Caroline go to the LGBTQ support group?\",\"answer\":\"7 May 2023\",\"evidence\":[\"D1:3\"],\"category\":2}",
"{\"question\":\"When did Melanie paint a sunrise?\",\"answer\":2022,\"evidence\":[\"D1:12\"],\"category\":2}",
"{\"question\":\"What fields would Caroline be likely to pursue ... | {"speaker_a":"Caroline","speaker_b":"Melanie","session_1_date_time":"1:56 pm on 8 May, 2023","session_1":[{"speaker":"Caroline","dia_id":"D1:1","text":"Hey Mel! Good to see you! How have you been?"},{"speaker":"Melanie","dia_id":"D1:2","text":"Hey Caroline! Good to see you! I'm swamped with the kids & work. What's up w... | {"events_session_1":{"Caroline":["Caroline attends an LGBTQ support group for the first time."],"Melanie":[],"date":"8 May, 2023"},"events_session_2":{"Caroline":["Caroline is inspired by her supportive friends and mentors to start researching adoption agencies."],"Melanie":[],"date":"25 May, 2023"},"events_session_3":... | {"session_1_observation":{"Caroline":[["Caroline attended an LGBTQ support group recently and found the transgender stories inspiring.","D1:3"],["The support group has made Caroline feel accepted and given her courage to embrace herself.","D1:7"],["Caroline is planning to continue her education and explore career optio... | {"session_1_summary":"Caroline and Melanie had a conversation on 8 May 2023 at 1:56 pm. Caroline mentioned that she attended an LGBTQ support group and was inspired by the transgender stories she heard. The support group made her feel accepted and gave her the courage to embrace herself. Caroline plans to continue her ... | conv-26 |
["{\"question\":\"When Jon has lost his job as a banker?\",\"answer\":\"19 January, 2023\",\"evidenc(...TRUNCATED) | "{\"speaker_a\":\"Jon\",\"speaker_b\":\"Gina\",\"session_1_date_time\":\"4:04 pm on 20 January, 2023(...TRUNCATED) | "{\"events_session_1\":{\"Jon\":[\"Jon loses his job as a banker.\",\"Jon begins planning for his ow(...TRUNCATED) | "{\"session_1_observation\":{\"Gina\":[[\"Gina lost her job at Door Dash during the month of the con(...TRUNCATED) | "{\"session_1_summary\":\"Gina and Jon met at 4:04 pm on 20 January, 2023. Jon lost his job as a ban(...TRUNCATED) | conv-30 |
["{\"question\":\"Who did Maria have dinner with on May 3, 2023?\",\"answer\":\"her mother\",\"evide(...TRUNCATED) | "{\"speaker_a\":\"John\",\"speaker_b\":\"Maria\",\"session_1_date_time\":\"11:01 am on 17 December, (...TRUNCATED) | "{\"events_session_1\":{\"John\":[\"John, his wife and their four kids, take a road trip together.\"(...TRUNCATED) | "{\"session_1_observation\":{\"Maria\":[[\"Maria volunteers at a homeless shelter and recently start(...TRUNCATED) | "{\"session_1_summary\":\"Maria and John met at 11:01 am on 17 December, 2022. Maria mentioned she h(...TRUNCATED) | conv-41 |
["{\"question\":\"Is it likely that Nate has friends besides Joanna?\",\"answer\":\"Yesteammates on (...TRUNCATED) | "{\"speaker_a\":\"Joanna\",\"speaker_b\":\"Nate\",\"session_1_date_time\":\"7:31 pm on 21 January, 2(...TRUNCATED) | "{\"events_session_1\":{\"Joanna\":[],\"Nate\":[\"Global Offensive with a team.\"],\"date\":\"21 Jan(...TRUNCATED) | "{\"session_1_observation\":{\"Nate\":[[\"Nate won his first video game tournament playing a team sh(...TRUNCATED) | "{\"session_1_summary\":\"Nate and Joanna caught up at 7:31 pm on 21 January, 2022. Nate won his fir(...TRUNCATED) | conv-42 |
["{\"question\":\"what are John's goals with regards to his basketball career?\",\"answer\":\"improv(...TRUNCATED) | "{\"speaker_a\":\"Tim\",\"speaker_b\":\"John\",\"session_1_date_time\":\"7:48 pm on 21 May, 2023\",\(...TRUNCATED) | "{\"events_session_1\":{\"Tim\":[\"Tim contacts his Harry Potter fan friend to discuss potential col(...TRUNCATED) | "{\"session_1_observation\":{\"John\":[[\"John signed with the Minnesota Wolves for the upcoming sea(...TRUNCATED) | "{\"session_1_summary\":\"John and Tim had a conversation at 7:48 pm on 21 May, 2023. Tim is working(...TRUNCATED) | conv-43 |
["{\"question\":\"Which year did Audrey adopt the first three of her dogs?\",\"answer\":\"2020\",\"e(...TRUNCATED) | "{\"speaker_a\":\"Audrey\",\"speaker_b\":\"Andrew\",\"session_1_date_time\":\"1:10 pm on 27 March, 2(...TRUNCATED) | "{\"events_session_1\":{\"Audrey\":[\"Audrey purchases three collars and tags for her dogs Pepper, P(...TRUNCATED) | "{\"session_1_observation\":{\"Audrey\":[[\"Audrey got new collars and tags for her dogs.\",\"D1:3\"(...TRUNCATED) | "{\"session_1_summary\":\"At 1:10 pm on 27 March 2023, Audrey and Andrew caught up. Andrew shared ab(...TRUNCATED) | conv-44 |
["{\"question\":\"What are John's suspected health problems?\",\"answer\":\"Obesity\",\"evidence\":[(...TRUNCATED) | "{\"speaker_a\":\"James\",\"speaker_b\":\"John\",\"session_1_date_time\":\"3:47 pm on 17 March, 2022(...TRUNCATED) | "{\"events_session_1\":{\"James\":[\"creating a game and telling a story through it.\",\"James propo(...TRUNCATED) | "{\"session_1_observation\":{\"John\":[[\"John signed up for a programming class to refresh his skil(...TRUNCATED) | "{\"session_1_summary\":\"John and James spoke at 3:47 pm on 17th March 2022. John asked James about(...TRUNCATED) | conv-47 |
["{\"question\":\"What kind of project was Jolene working on in the beginning of January 2023?\",\"a(...TRUNCATED) | "{\"speaker_a\":\"Deborah\",\"speaker_b\":\"Jolene\",\"session_1_date_time\":\"4:06 pm on 23 January(...TRUNCATED) | "{\"events_session_1\":{\"Deborah\":[\"Deborah visits her mother's old home while reflecting on her (...TRUNCATED) | "{\"session_1_observation\":{\"Deborah\":[[\"Deborah finished an electrical engineering project last(...TRUNCATED) | "{\"session_1_summary\":\"Deborah and Jolene met at 4:06 pm on 23 January, 2023. They discussed thei(...TRUNCATED) | conv-48 |
["{\"question\":\"What kind of car does Evan drive?\",\"answer\":\"Prius\",\"evidence\":[\"D1:2\",\"(...TRUNCATED) | "{\"speaker_a\":\"Evan\",\"speaker_b\":\"Sam\",\"session_1_date_time\":\"1:47 pm on 18 May, 2023\",\(...TRUNCATED) | "{\"events_session_1\":{\"Evan\":[\"Evan buys a new Prius after he repairs and sells his old Prius w(...TRUNCATED) | "{\"session_1_observation\":{\"Sam\":[[\"Sam went hiking with his dad when he was ten, which was a s(...TRUNCATED) | "{\"session_1_summary\":\"Sam and Evan met at 1:47 pm on 18 May, 2023. Evan mentioned his recent tri(...TRUNCATED) | conv-49 |
["{\"question\":\"When did Calvin first travel to Tokyo?\",\"answer\":\"between 26 March and 20 Apri(...TRUNCATED) | "{\"speaker_a\":\"Calvin\",\"speaker_b\":\"Dave\",\"session_1_date_time\":\"11:53 am on 23 March, 20(...TRUNCATED) | "{\"events_session_1\":{\"Calvin\":[\"Calvin purchases a mansion in Japan and plans to stay there fo(...TRUNCATED) | "{\"session_1_observation\":{\"Calvin\":[[\"Calvin recently had a big life change and now has a new (...TRUNCATED) | "{\"session_1_summary\":\"Calvin and Dave met at 11:53 am on 23 March, 2023. Dave attended a car eve(...TRUNCATED) | conv-50 |
End of preview.
中文记忆评测数据集合（Chinese Memory-Eval Datasets）
面向长对话记忆系统评测的两个中文数据集，均由英文原版全量翻译（模型 glm-5.3-flash + 全局术语表统一人名地名译法），结构与源数据逐字段对齐，100% 全量校验（结构 0 错、无漏译、无中英混杂）。
| 子目录 | 数据集 | 规模 |
|---|---|---|
longmemeval-s/ |
LongMemEval-S 中文版（源：xiaowu0162/longmemeval-cleaned，Wu et al., arXiv:2410.10813） | 470 题（30 题因敏感内容无法翻译而剔除，清单见 pruned_questions.json），194MB |
locomo/ |
LoCoMo 中文版（源：snap-research/locomo，Maharana et al., ACL 2024） | 10 段对话 / 5,882 条消息 / 1,986 条 QA，2.0MB |
下载
from huggingface_hub import snapshot_download
snapshot_download("justis-xu/memory-eval-zh", repo_type="dataset", local_dir="./memory-eval-zh")
# 只下某个子集：
snapshot_download("justis-xu/memory-eval-zh", repo_type="dataset",
local_dir="./locomo", allow_patterns=["locomo/*"])
或命令行：hf download justis-xu/memory-eval-zh --repo-type dataset --local-dir ./memory-eval-zh
各子目录的 README 里有字段说明、翻译口径（保留原文的类别）与加载示例。
- Downloads last month
- -