What's Changed
* [Bug-Fix] fix attention head intervention for multiple models by Bakser in https://github.com/stanfordnlp/pyvene/pull/159
* [P2] Add Sparse Autoencoder Interventions by explanare in https://github.com/stanfordnlp/pyvene/pull/164
* [P0] Revert back the kwargs argument for intervention init by frankaging in https://github.com/stanfordnlp/pyvene/pull/166
* [P0] First infrastructure change to support NDIF backend by frankaging in https://github.com/stanfordnlp/pyvene/pull/167
* [P2] Save/load trainable params in `IntervenableBase` methods by aryamanarora in https://github.com/stanfordnlp/pyvene/pull/153
* [P1] Start docs website by aryamanarora in https://github.com/stanfordnlp/pyvene/pull/169
* device of unit_locations should follow tensor_input by aryopg in https://github.com/stanfordnlp/pyvene/pull/171
* [Minor] Fix use_cache flag propagation by frankaging in https://github.com/stanfordnlp/pyvene/pull/172
* [Minor] Fix the base model use_cache arg passing by frankaging in https://github.com/stanfordnlp/pyvene/pull/173
* [Minor] Update `use_cache` intervenable_base.py by frankaging in https://github.com/stanfordnlp/pyvene/pull/175
* [Minor] Start to support generic intervention output, and adaptor-like tuning by frankaging in https://github.com/stanfordnlp/pyvene/pull/177
* add support for dict outputs (select the first value) by leogagnon in https://github.com/stanfordnlp/pyvene/pull/178
* [Minor] Update the huggingface-hub version by frankaging in https://github.com/stanfordnlp/pyvene/pull/179
New Contributors
* Bakser made their first contribution in https://github.com/stanfordnlp/pyvene/pull/159
* explanare made their first contribution in https://github.com/stanfordnlp/pyvene/pull/164
* aryopg made their first contribution in https://github.com/stanfordnlp/pyvene/pull/171
* leogagnon made their first contribution in https://github.com/stanfordnlp/pyvene/pull/178
**Full Changelog**: https://github.com/stanfordnlp/pyvene/compare/v0.1.2...v0.1.3