datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
bento_ur7e_v1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 60,
"features": {
"action": {
"dtype": "float32",
"names": [
"joint_0.pos",
"joint_1.pos",
"joint_2.pos",
"joint_3.pos",
"joint_4.pos",
"joint_5.pos",
"gripper.pos"
],
"shape": [… See the full description on the dataset page: https://huggingface.co/datasets/polarisai-robots/bento_ur7e_v1.maxrl_qwen3_4B_base_polaris_rollouts
MaxRL Qwen3-4B-Base training rollouts (POLARIS math prompts)
Every training rollout from an online RL run, with exact token ids, sampling
log-probs, and raw rewards — usable as a replay buffer to study off-policy RL
for LLM reasoning completely offline.
The run: Qwen3-4B-Base trained with the maxRL advantage estimator
(A = (r - mean)/(mean + eps), group mean over 16 rollouts per prompt;
maxRL paper) and a pure REINFORCE loss
(L = -A * log pi; no importance ratio, no clipping, no… See the full description on the dataset page: https://huggingface.co/datasets/ftajwar/maxrl_qwen3_4B_base_polaris_rollouts.qwen3-1.7b-polaris-fp8-rollouts-20260912
Qwen3-1.7B Polaris FP8 rollouts
Snapshot of three FP8 rollout datasets taken on 2026-09-12. The model is Qwen3-1.7B-Base. Original Parquet files and attempt, file, and checkpoint-lineage metadata are preserved without rewriting.
Configuration
Training batches
Training responses
Validation responses
Total bytes
maxrl_strict
101
827392
144320
6654288117
maxrl_permissive
107
876544
144320
8168064932
dppo
126
1032192
173184
6733682534
Provenance and… See the full description on the dataset page: https://huggingface.co/datasets/steviel/qwen3-1.7b-polaris-fp8-rollouts-20260912.rloo_qwen3_1p7B_base_polaris_rolloutspolaris-ecir-v1
ECIR v1
ECIR is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, lter, wikitables, and
wtr.
It holds 2,100 tables published on the US government open data portal and 12 keyword queries over
them. For each query–table pair, a person decided how well that table answers that query and gave it
a score; those scores are the relevance judgments, and they live in qrels.csv. Given a query, a
system ranks the 2,100… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-ecir-v1.polaris-wikitables-v2
WikiTables v2
WikiTables is one of six datasets in Polaris: Learning to Generate Table Descriptions from
Retrieval Feedback, alongside aw, arctic, lter, ecir,
and wtr.
It holds 3,361 tables scraped from Wikipedia articles and 57 keyword queries over them. For each
query–table pair, a person scored how well that table answers that query; those scores are the
relevance judgments, and they live in qrels.csv.
The tables have no names. What describes a table is its column names and… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-wikitables-v2.polaris-acemath-gemini-rubrics-v2bento_v2_openarmThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 30,
"features": {
"action": {
"dtype": "float32",
"names": [
"joint_1.pos",
"joint_1.vel",
"joint_1.torque",
"joint_2.pos",
"joint_2.vel",
"joint_2.torque",
"joint_3.pos",
"joint_3.vel"… See the full description on the dataset page: https://huggingface.co/datasets/polarisai-robots/bento_v2_openarm.polaris-lter-v1
LTER v1
LTER is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, ecir, wikitables, and
wtr.
It holds 2,015 tables from the top-downloaded collections of the Long-Term Ecological Research sites
in the Environmental Data Initiative (EDI) — bird surveys, forest phenology, reef colonisation,
cattle records — and 15 keyword queries over them. For each query–table pair, a person decided
whether that table answers… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-lter-v1.pretrain_polarispolaris-53k-repaired
POLARIS-53K, label-repaired
49,289 of the 53,291 rows in
POLARIS-Project/Polaris-Dataset-53K,
with 4,580 stored answers corrected and 4,002 rows removed as unrepairable.
Measurements on the source set put its bad-label rate at roughly 15.9%
[14.3, 17.6] (two independent detectors agreeing on a 2,000-row sample).
Mislabelled rows are not uniformly distributed: they concentrate in the problems
models fail, which is exactly where a difficulty-calibration pipeline looks.… See the full description on the dataset page: https://huggingface.co/datasets/joanvelja/polaris-53k-repaired.grpo_qwen3_1p7B_base_polaris_rolloutspolaris-aw-v2
AW v2
AW is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside arctic, lter, ecir, wikitables, and
wtr.
It holds 96 tables from a version of AdventureWorks whose column names are cryptic — BusEntId,
STRGUID, JobTtl — and 15 keyword queries over them. For each query–table pair, a person decided
whether that table answers that query; those decisions are the relevance judgments, and they live in
qrels.csv.
Each Polaris dataset… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-aw-v2.bento_ur7e_v1_20260908_104320This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 60,
"features": {
"action": {
"dtype": "float32",
"names": [
"joint_0.pos",
"joint_1.pos",
"joint_2.pos",
"joint_3.pos",
"joint_4.pos",
"joint_5.pos",
"gripper.pos"
],
"shape": [… See the full description on the dataset page: https://huggingface.co/datasets/polarisai-robots/bento_ur7e_v1_20260908_104320.polaris-ecir-v2
ECIR v2
ECIR is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, lter, wikitables, and
wtr.
It holds 2,100 tables published on the US government open data portal and 12 keyword queries over
them. For each query–table pair, a person decided how well that table answers that query and gave it
a score; those scores are the relevance judgments, and they live in qrels.csv. Given a query, a
system ranks the 2,100… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-ecir-v2.maxrl_qwen3_1p7B_base_polaris_rolloutsPOLARIS-Gov
POLARIS-Gov: Political Orientation Longitudinal Analysis Resource for International States
Dataset Statistics
Countries Tracked: 171 Sovereign Nations (160 in effective modeling universe)
Observations: 3,373 country-year rows
Temporal Coverage: 2000–2019
Treatment Variable: ideology_score (Unified ideological score of governing coalitions)
Outcome Variables: GDP growth, GDP per capita growth
Control Variables: 18 World Bank Economic Indicators
Panel Type:… See the full description on the dataset page: https://huggingface.co/datasets/Maybe-Heisenberg-07/POLARIS-Gov.polaris-wtr-v1
WTR v1
WTR is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, lter, ecir, and
wikitables.
It holds 4,634 tables crawled from web pages and 60 keyword queries over them. For each query–table
pair, a person scored how well that table answers that query; those scores are the relevance
judgments, and they live in qrels.csv.
The tables have no names. What describes a table is its column names and the text around… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-wtr-v1.polaris-arctic-v1
Arctic v1
Arctic is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, lter, ecir, wikitables, and
wtr.
It holds 251 tables sampled at random from the Environmental Data Initiative (EDI), a repository of
long-term ecological research data — lake water temperature, soil chemistry, rainfall, coral
taxonomy — and 20 keyword queries over them. For each query–table pair, a person decided whether that
table answers that… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-arctic-v1.Polaris-hard-w-solutions-24209
Polaris-Hard-w-Solutions
24,209 hard competition-math problems (the hardest difficulty bands of the
Polaris dataset) paired with
two verified solutions each: a full original solution and a concise summarized solution. Every
retained problem has a machine-verifiable final answer, every solution's boxed answer grades correct
against the reference (sympy-based grading), and the summarized solutions have additionally been put
through a reasoning-rigor pass (see step 5 below).
This… See the full description on the dataset page: https://huggingface.co/datasets/JWei05/Polaris-hard-w-solutions-24209.prm-sft-polaris-mc
PRM-SFT Polaris MC — Monte-Carlo value (unfiltered)
Soft-target process reward model (PRM) training data with per-prefix
Monte-Carlo values, in the style of Math-Shepherd.
Each row is one prefix of a reasoning trace on a Polaris math problem, labeled with
V(prefix)=P(correct∣prefix)=#correct continuations#continuationsV(\text{prefix}) = P(\text{correct} \mid \text{prefix}) = \frac{\#\text{correct… See the full description on the dataset page: https://huggingface.co/datasets/asingh15/prm-sft-polaris-mc.polaris-wikitables-v1
WikiTables v1
WikiTables is one of six datasets in Polaris: Learning to Generate Table Descriptions from
Retrieval Feedback, alongside aw, arctic, lter, ecir,
and wtr.
It holds 3,361 tables scraped from Wikipedia articles and 57 keyword queries over them. For each
query–table pair, a person scored how well that table answers that query; those scores are the
relevance judgments, and they live in qrels.csv.
The tables have no names. What describes a table is its column names and… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-wikitables-v1.polaris_rose_rollouts_olmo3-7b_from_qwen3-30b-a3b_cutoff4096_240steps
Cross-tokenizer ROSE rollouts — Olmo-3-7B-Think-SFT ← Qwen3-30B-A3B-Thinking-2507
Every assembled row of a complete 240-step online-ROSE run: 61,440 rows, the teacher's
actual continuation for each, and the token accounting behind it.
The student writes a 4096-token prefix in its own vocabulary (100278). That prefix is
decoded to text, the teacher is shown it under its own chat template, and the teacher's
reply comes back as text and is tokenised into the student's vocabulary.… See the full description on the dataset page: https://huggingface.co/datasets/SeanWang0027/polaris_rose_rollouts_olmo3-7b_from_qwen3-30b-a3b_cutoff4096_240steps.polaris-arctic-v2
Arctic v2
Arctic is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, lter, ecir, wikitables, and
wtr.
It holds 251 tables sampled at random from the Environmental Data Initiative (EDI), a repository of
long-term ecological research data — lake water temperature, soil chemistry, rainfall, coral
taxonomy — and 20 keyword queries over them. For each query–table pair, a person decided whether
that
table answers that… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-arctic-v2.polaris-wtr-v2
WTR v2
WTR is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, lter, ecir, and
wikitables.
It holds 4,634 tables crawled from web pages and 60 keyword queries over them. For each query–table
pair, a person scored how well that table answers that query; those scores are the relevance
judgments, and they live in qrels.csv.
The tables have no names. What describes a table is its column names and the text around… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-wtr-v2.polaris-lter-v2
LTER v2
LTER is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside aw, arctic, ecir, wikitables, and
wtr.
It holds 2,015 tables from the top-downloaded collections of the Long-Term Ecological Research sites
in the Environmental Data Initiative (EDI) — bird surveys, forest phenology, reef colonisation,
cattle records — and 15 keyword queries over them. For each query–table pair, a person decided
whether that table answers… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-lter-v2.ROSE-polaris-popepolaris_molmoact2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"fps": 20,
"features": {
"observation.images.image": {
"dtype": "video",
"shape": [
720,
1280,
3
],
"names": [
"height",
"width",
"channels"
],
"info": {
"video.height":… See the full description on the dataset page: https://huggingface.co/datasets/Makolon0321/polaris_molmoact2.polaris-aw-v1
AW v1
AW is one of six datasets in Polaris: Learning to Generate Table Descriptions from Retrieval
Feedback, alongside arctic, lter, ecir, wikitables, and
wtr.
It holds 96 tables from a version of AdventureWorks whose column names are cryptic — BusEntId,
STRGUID, JobTtl — and 15 keyword queries over them. For each query–table pair, a person decided
whether that table answers that query; those decisions are the relevance judgments, and they live in
qrels.csv.
Each Polaris dataset… See the full description on the dataset page: https://huggingface.co/datasets/anhaidgroup/polaris-aw-v1.fixed-n-rb-offset-marginrl-qwen3-4b-base-polaris53k-offset512-token-mean-1epoch-rollouts
fixed_n_rb_offset_marginrl_Qwen3-4B-Base_polaris53k_offset512_token_mean_1epoch rollouts
This dataset contains one compressed JSONL shard for every completed training
step. The step and rollout_index columns uniquely locate a rollout within
this training run. Run metadata and per-step row counts are recorded in
rollout_manifest.json.