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: Couldn't cast array of type
struct<cmd: string, output: string, exit: int64, cwd: string, meta: struct<verb: string, linked: bool, query_src: string, availability: struct<head: bool, tail: bool, stat: bool, find: bool, grep: bool>, skipped_verbs: list<item: null>, lexicon_version: string, pool: string, intended_empty: bool, hit: bool, band: string, trunc_stored: bool, intended_miss: bool, flags: string, nav: string, k: int64>>
to
{'cmd': Value('string'), 'output': Value('string'), 'exit': Value('int64'), 'cwd': Value('string')}
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_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 2312, in cast_table_to_schema
cast_array_to_feature(
~~~~~~~~~~~~~~~~~~~~~^
table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
feature,
^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1861, in wrapper
return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
~~~~^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2118, in cast_array_to_feature
casted_array_values = _c(array.values, feature.feature)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1863, in wrapper
return func(array, *args, **kwargs)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2158, in cast_array_to_feature
raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
TypeError: Couldn't cast array of type
struct<cmd: string, output: string, exit: int64, cwd: string, meta: struct<verb: string, linked: bool, query_src: string, availability: struct<head: bool, tail: bool, stat: bool, find: bool, grep: bool>, skipped_verbs: list<item: null>, lexicon_version: string, pool: string, intended_empty: bool, hit: bool, band: string, trunc_stored: bool, intended_miss: bool, flags: string, nav: string, k: int64>>
to
{'cmd': Value('string'), 'output': Value('string'), 'exit': Value('int64'), 'cwd': Value('string')}
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 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 1880, 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.
image string | system_id string | steps list |
|---|---|---|
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -o",
"output": "GNU/Linux",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/profile",
"output": "# /etc/profile: system-wide .profile file for the Bourne shell (sh(1))\n# and Bourne compatible shells (bash(1), ksh(1), ash(1), ...).\n\nif [ \"$(id -u)\" -eq 0 ]; then\n PATH=\... |
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -n",
"output": "b42097a46cfc",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/hostname",
"output": "b42097a46cfc",
"exit": 0,
"cwd": "/"
},
{
"cmd": "ls -la",
"output": "total 56\ndrwxr-xr-x 1 root root 4096 Jul 18 21:08 .\ndrwxr-xr-x 1 root root 4096... |
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -r",
"output": "6.6.65-0-virt",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/os-release",
"output": "PRETTY_NAME=\"Debian GNU/Linux 13 (trixie)\"\nNAME=\"Debian GNU/Linux\"\nVERSION_ID=\"13\"\nVERSION=\"13 (trixie)\"\nVERSION_CODENAME=trixie\nDEBIAN_VERSION_FULL=13.6\nID=d... |
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -n",
"output": "b42097a46cfc",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/passwd",
"output": "root:x:0:0:root:/root:/bin/bash\ndaemon:x:1:1:daemon:/usr/sbin:/usr/sbin/nologin\nbin:x:2:2:bin:/bin:/usr/sbin/nologin\nsys:x:3:3:sys:/dev:/usr/sbin/nologin\nsync:x:4:65534:sync... |
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -m",
"output": "aarch64",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/resolv.conf",
"output": "# Generated by Docker Engine.\n# This file can be edited; Docker Engine will not make further changes once it\n# has been modified.\n\nnameserver 192.168.5.3\n\n# Based on host ... |
debian:stable-slim | Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 13 (trixie)"
NAME="Debian GNU/Linux" | [
{
"cmd": "uname -sr",
"output": "Linux 6.6.65-0-virt",
"exit": 0,
"cwd": "/"
},
{
"cmd": "cat /etc/shells",
"output": "# /etc/shells: valid login shells\n/bin/sh\n/usr/bin/sh\n/bin/bash\n/usr/bin/bash\n/bin/rbash\n/usr/bin/rbash\n/usr/bin/dash",
"exit": 0,
"cwd": "/"
},
{
... |
debian:stable-slim | "Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linu(...TRUNCATED) | [{"cmd":"uname -m","output":"aarch64","exit":0,"cwd":"/"},{"cmd":"cat /etc/passwd","output":"root:x:(...TRUNCATED) |
debian:stable-slim | "Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linu(...TRUNCATED) | [{"cmd":"uname -s","output":"Linux","exit":0,"cwd":"/"},{"cmd":"cat /proc/version","output":"Linux v(...TRUNCATED) |
debian:stable-slim | "Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linu(...TRUNCATED) | [{"cmd":"uname -n","output":"b42097a46cfc","exit":0,"cwd":"/"},{"cmd":"cat /etc/hosts","output":"127(...TRUNCATED) |
debian:stable-slim | "Linux b42097a46cfc 6.6.65-0-virt #1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04 aarch64 GNU/Linu(...TRUNCATED) | [{"cmd":"uname -v","output":"#1-Alpine SMP PREEMPT_DYNAMIC 2024-12-12 12:13:04","exit":0,"cwd":"/"},(...TRUNCATED) |
End of preview.
terminal-jepa dockerfs
Shell exploration sequences over stock public Docker images (alpine/ubuntu/debian/python/redis/nginx/postgres/node train; fedora/rockylinux/mariadb/httpd held-out val), collected for the terminal-JEPA world-model research program. Each sequence interleaves navigational commands (uname/cat/ls/cd) with their observed outputs inside the container.
dockerfs/— raw collected sequences (train.jsonl,val.jsonl) +summary.json. Fully regenerable: the collection RNG is seeded (realenv/collect_docker.py, seed 0, 400 seqs/image, len 16).dockerfs-e5/— the same sequences re-encoded with frozenintfloat/e5-base-v2(emb-seq-*.ptcaches consumed by the scoring harness).
dockerfs3 (v3.0) — the mutation dynamical world (2026-07-23)
The action space grew from read-only navigation to a mutating, dynamical filesystem: rm/mv/ln/echo-redirect/mkdir/touch (state mutation), pipes/redirection/conditionals (composition), and a ps/kill/after process automaton — over 12 digest-pinned images. 201,160 steps (dockerfs3/) + a 67,130-step train-only ablate arm (dockerfs3-ablate/, mutation/composition/time OFF) for a paired causal comparison; e5 re-encodings in dockerfs3-e5/ + dockerfs3-ablate-e5/ (the latter carry sst-val.pt/wtm-val.pt — the symbolic-shell-state-tracker baseline arms). Every step is folded by a symbolic shell-state tracker (realenv/shell_state.py) that doubles as the honesty baseline; the raw jsonl is byte-identical across twin mints (store-time determinism canonicalization), and classes_sha pins the frozen fitness class table into every root's summary.json.
Honest fitness scope: the mint-scale measurement showed most of the new breadth is not fitness-bearing — head/tail/stat are image-memorizable (a fixed read of a fixed file is a constant; no-history-predictability ≥ threshold → echo/const), grep/composition are coverage-starved on the held-out splits, and the process automaton is report-only. The fitness surface is 5 cells — ls/cat (native + mutated), find (native) — where the genuine advance over v2 is the mutated-scope cells: predicting reads of mutation-altered paths requires tracking the dynamical filesystem. The full 22-verb data is published (the honesty gating is a scoring-layer decision recorded in the prereg); the read-verb/composition/process breadth is available for study and is report-only pending a v3.1 with more held-out coverage.
All content is machine output of unmodified public container images; nothing user- or host-specific.
- Downloads last month
- 10