datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SAEVerbalizer-Data
SAEVerbalizer Data
Evaluation data for SAEVerbalizer: Generating Explanations for Sparse
Autoencoder Features via Representation Verbalization.
This release contains the paper's three evaluation sets for both the 27B
verbalizer and the 1B-to-27B adapter-based verbalization system. Training data
may be added in a later release.
Inference and Reference Agreement evaluation code is available in
THU-KEG/SAEVerbalizer.
Data collections
Collection
Feature space… See the full description on the dataset page: https://huggingface.co/datasets/THU-KEG/SAEVerbalizer-Data.sa-evt2-01-cable11This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 3,
"total_frames": 1305,
"total_tasks": 1,
"total_videos": 9,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-cable11.llavasae_obliec100k_SAEVsa-evt2-01-test8This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 251,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test8.sa-evt2-01-cable8This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 3,
"total_frames": 1377,
"total_tasks": 1,
"total_videos": 9,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-cable8.sa-evt2-01-cable9This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 3,
"total_frames": 1789,
"total_tasks": 1,
"total_videos": 9,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-cable9.sa-evt2-01-test-fixqp1This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 3,
"total_frames": 1315,
"total_tasks": 1,
"total_videos": 9,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test-fixqp1.sa-evt2-01-test2This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 2,
"total_frames": 490,
"total_tasks": 1,
"total_videos": 6,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:2"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test2.sa-evt2-01-test3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 121,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test3.sa-evt2-01-cable5This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 151,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-cable5.sa-evt2-01-cable7This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 502,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-cable7.sa-evt2-01-test6This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 244,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test6.sa-evt2-01-test24This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 254,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test24.sa-evt2-01-test23This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "eyou_ft7_follower",
"total_episodes": 1,
"total_frames": 78,
"total_tasks": 1,
"total_videos": 3,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/1g0rrr/sa-evt2-01-test23.SAE-V
SAE-V
Repository for SAE-V, including 2 sparse autoencoder (SAE) and 3 sparse autoencoder with Vision (SAE-V). See each model folders for more information.
1.Training Parameter
All 5 models training paramters are list below:
Hyper-parameters
SAE and SAE-V of LLaVA-NeXT/Mistral
SAE and SAE-V of Chameleon/Anole
Training Parameters
total training steps
30000
30000
batch size… See the full description on the dataset page: https://huggingface.co/datasets/saev-2025/SAE-V.