Official checkpoints for paper DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning.
HAO BAI
JackBAI
AI & ML interests
Representation learning, language models.
Recent Activity
updated
a model
about 2 months ago
JackBAI/CRATE-GPT-12L-Pile-600000steps
Organizations
Collections
1
models
15
JackBAI/CRATE-GPT-12L-Pile-600000steps
Updated
JackBAI/webshop-off2on-filteredbc
Updated
JackBAI/general-off2on-filteredbc
Updated
JackBAI/general-off2on-digirl
Updated
•
1
JackBAI/webshop-off2on-digirl
Updated
JackBAI/crate-3l-l0-sae-1x
Updated
JackBAI/crate-1l-l0-sae-1x
Updated
JackBAI/gpt-1l-l0-sae-1x
Updated
JackBAI/crate-3l-sae-16x
Updated
JackBAI/crate-1l-sae-16x
Updated