Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code: DatasetGenerationError
Exception: TypeError
Message: int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1539, in _prepare_split_single
for key, record in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 621, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 131, in _generate_examples
for example_idx, example in enumerate(self._get_pipeline_from_tar(tar_path, tar_iterator)):
~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 32, in _get_pipeline_from_tar
for filename, f in tar_iterator:
^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/track.py", line 49, in __iter__
for x in self.generator(*self.args):
~~~~~~~~~~~~~~^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 1419, in _iter_from_urlpath
with xopen(urlpath, "rb", download_config=download_config, block_size=0) as f:
~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/file_utils.py", line 976, in xopen
file_obj = fs.open(paths[0], mode)
File "<string>", line 3, in open
File "/usr/local/lib/python3.14/unittest/mock.py", line 1176, in __call__
return self._mock_call(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/unittest/mock.py", line 1180, in _mock_call
return self._execute_mock_call(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/unittest/mock.py", line 1247, in _execute_mock_call
result = effect(*args, **kwargs)
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 794, in wrapped
tracker.files[urlpath] = {"read": 0, "size": int(f.size)}
~~~^^^^^^^^
TypeError: int() argument must be a string, a bytes-like object or a real number, not 'NoneType'
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/parquet_and_info.py", line 1379, 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 956, 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 1401, 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 1579, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
log unknown | __key__ string | __url__ string |
|---|---|---|
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_noinfo_501 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
104,
105,
112,
112,
111,
114,
97,
103,
58,
32,
53,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
95,
... | ./hipporag | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_noinfo_701 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_701 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
109,
97,
103,
101,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
95,
115,
116,
101,
112,
... | ./mage_noinfo | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
104,
121,
98,
114,
105,
100,
51,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
95,
115,
1... | ./hybrid3 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_noinfo_401 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_101 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_401 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
115,
105,
109,
112,
108,
101,
109,
101,
109,
58,
32,
53,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
53,
59,
32,
109,
97,
120,
... | ./simplemem_w5 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_noinfo_101 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_501 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_noinfo_801 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
109,
97,
103,
101,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
95,
115,
116,
101,
112,
... | ./mage | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
109,
97,
103,
101,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
95,
115,
116,
101,
112,
... | ./mage_oracle | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
108,
111,
110,
103,
95,
99,
111,
110,
116,
101,
120,
116,
58,
32,
48,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
49,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
49,
59,
32,
... | ./long_context_801 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
98,
101,
110,
99,
104,
53,
47,
115,
105,
109,
112,
108,
101,
109,
101,
109,
58,
32,
53,
32,
112,
101,
110,
100,
105,
110,
103,
32,
111,
102,
32,
53,
59,
32,
119,
111,
114,
107,
101,
114,
115,
32,
52,
59,
32,
109,
97,
120,
... | ./simplemem | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/_logs/run.tar.zst |
[
65,
103,
101,
110,
116,
32,
98,
101,
110,
99,
104,
53,
95,
100,
101,
115,
107,
95,
97,
103,
101,
110,
116,
32,
114,
101,
99,
101,
105,
118,
101,
115,
32,
116,
104,
101,
32,
102,
111,
108,
108,
111,
119,
105,
110,
103,
32,
... | ./task_801_subtask_10 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/bench5/hipporag/run.tar.zst |
[
65,
103,
101,
110,
116,
32,
98,
101,
110,
99,
104,
53,
95,
100,
101,
115,
107,
95,
97,
103,
101,
110,
116,
32,
114,
101,
99,
101,
105,
118,
101,
115,
32,
116,
104,
101,
32,
102,
111,
108,
108,
111,
119,
105,
110,
103,
32,
... | ./task_701_subtask_13 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/bench5/hipporag/run.tar.zst |
[
65,
103,
101,
110,
116,
32,
98,
101,
110,
99,
104,
53,
95,
100,
101,
115,
107,
95,
97,
103,
101,
110,
116,
32,
114,
101,
99,
101,
105,
118,
101,
115,
32,
116,
104,
101,
32,
102,
111,
108,
108,
111,
119,
105,
110,
103,
32,
... | ./task_101_subtask_0 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/bench5/hipporag/run.tar.zst |
[
65,
103,
101,
110,
116,
32,
98,
101,
110,
99,
104,
53,
95,
100,
101,
115,
107,
95,
97,
103,
101,
110,
116,
32,
114,
101,
99,
101,
105,
118,
101,
115,
32,
116,
104,
101,
32,
102,
111,
108,
108,
111,
119,
105,
110,
103,
32,
... | ./task_501_subtask_17 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/bench5/hipporag/run.tar.zst |
[
65,
103,
101,
110,
116,
32,
98,
101,
110,
99,
104,
53,
95,
100,
101,
115,
107,
95,
97,
103,
101,
110,
116,
32,
114,
101,
99,
101,
105,
118,
101,
115,
32,
116,
104,
101,
32,
102,
111,
108,
108,
111,
119,
105,
110,
103,
32,
... | ./task_801_subtask_27 | hf://datasets/eason-hk-barcelona/agentmemory-bench5@ba34691e5f316dfc1d5d606a301aa894eecf30d6/trajectories/results_bench5/long_v3b/iter1/bench5/hipporag/run.tar.zst |
End of preview.
AgentMemory bench5：数据和运行轨迹
bench5 是一个合成的"运营台值班" benchmark，测长程 agent 能不能在需要时想起早先顺带见过的事实：
- 第一天：agent 处理别的工单时，顺带看到一条事实，比如电话里的一句闲聊、长表里的一行、确认包里的一个条款。
- 最后一天：某张工单的决定恰好取决于这条事实。
一共五道题（t1g、t4、t5、t7、t8），每题 6 个 test 种子和 3 个 catch 种子，分两档：
- 短档：5 个 session，历史放得进 131K 窗口；
- 长档：32–50 个 session，约 26 万 token，是窗口的两倍。
代码、规格和报告都在 GitHub 仓库 eason-hk-barcelona/AgentMemory：
| 内容 | 位置 |
|---|---|
| bench5 代码和规格 | benchmarks/bench5/（SPEC.md） |
| 基线结果报告 | docs/reports/bench5 最终结果：超窗场景下的不显眼事实回忆.md |
| proactD 报告 | docs/reports/bench5 主动记忆检索 baseline：proactD.md |
目录
| 路径 | 内容 |
|---|---|
data/bench5.jsonl |
短档，45 个 bundle，md5 d18429b9901517a8de40100d0e2d366b |
data/bench5_long.jsonl |
长档，45 个 bundle，md5 f0ebb3fbcdc10b0e5a41ffcf411568a1 |
data/gate_report*.json |
静态门报告 |
trajectories/results_bench5/ |
基线运行，执行模型 Qwen3.6-35B-A3B |
trajectories/results_proact/ |
proactD 及其对照的运行 |
SHA256SUMS |
所有文件的校验和 |
基线运行（trajectories/results_bench5/）分这几组：
| 目录 | 内容 |
|---|---|
v3、v3_t8 |
短档主实验，iter1–6 是 test 种子，iter1–3_catch 是 catch 种子；协议 nothink |
v3_lean2_h3 |
短档 hybrid3 的 lean2 重跑 |
v3_thinkact、v3_nohand |
短档稳健性实验：动作调用开思考 / 去掉交接笔记 |
long_v3b |
长档；协议 lean2 |
方法包括 mage、hybrid3、SimpleMem、HippoRAG、long context，以及 oracle 和 noinfo 两个对照臂。
proactD 的运行（trajectories/results_proact/）分这几组：
| 目录 | 内容 |
|---|---|
short_full |
短档：D / B / L / A / C / D 随机证据 / D+TODAYSKIP，各 2 次 |
long_v2 |
长档，同一组变体 |
long_screen1 |
长档第一版代码 |
base_short_lean2 |
同协议的短档基线重跑 |
每个运行目录（.../bench5/<方法>/）下有三样东西：
results.jsonl：每题一行。sr是关键动作是否正确（严格判定），stats是各种机制计数。run_config*.json：运行参数，包括数据 md5、协议和 proact 开关。run.tar.zst：这个运行写下的全部文件：task_<id>_subtask_<k>.log：每个 session 的逐步动作和观察；task_<id>_token_usage.json：每次 LLM 调用的完整 prompt 和输出；task_<id>_bench5.json：ledger、判定、交接笔记、关键 session 每个检索块的原文；- proactD 还有
proact_<id>.json（每次触发的需求、作用域、送达行、拦截）、proact_<id>_archive.json和pr_events.jsonl。
各组的 _logs/run.tar.zst 是运行时的标准输出日志。
用法
# 只下数据，放到代码仓库里 bench5 读取的位置
hf download eason-hk-barcelona/agentmemory-bench5 --repo-type dataset --include "data/*" --local-dir AgentMemory/benchmarks/bench5
md5sum AgentMemory/benchmarks/bench5/data/bench5*.jsonl
# 下某一组轨迹并解开
hf download eason-hk-barcelona/agentmemory-bench5 --repo-type dataset --include "trajectories/results_proact/short_full/*" --local-dir bench5_hf
cd bench5_hf/trajectories/results_proact/short_full/bench5/proactD_r1 && zstd -dc run.tar.zst | tar -x
数据也可以用代码仓库里的 benchmarks/bench5/generate.py 重新生成（长档要 Qwen3.6-35B-A3B 的 tokenizer），md5 应与上表一致。
- Downloads last month
- 76