v0.0.1

v0.0.1 (2024-12-13)

Fix

• fix: experiment script typo (8a14c12)

• fix: build dataset (7ef7e8e)

Unknown

• add ci (#3)

• add ci, pre-commit

• fix ci

• update ci

• update ruff config

---

Co-authored-by: Daniel Tan <dtch1997@users.noreply.github.com> (33c8a2b)

• Update README.md (7ed9e3b)

• add arxiv badge (f302b40)

• update (15f3930)

• rename persona_generalization to steering_generalization (0ddd1ad)

• Update README.md (a17bfba)

• Update README.md (b566646)

• Update README.md (95295e6)

• Create README.md (fae2efb)

• Create README.md (9151b2a)

• Update README.md (0f25019)

• Refusal (#2)

• add refusal dataset

• wip refusal experiment

---

Co-authored-by: Daniel CH Tan <dtch1997@users.noreply.github.com> (893d67a)

• remove redundant script (3d4870f)

• refactor (ffe4666)

• defer formatting to apply_chat_template (53021b6)

• add persona generalization experiment (3d70293)

• delete results (8dac290)

• steering experiments (#1)

• wip steering experiment

• minimal working version done

• refactor

• add working layer sweep script

• working layer sweep

---

Co-authored-by: Daniel Tan <dtch1997@users.noreply.github.com> (b571672)

• add dataset builder (29783a3)

• update interfaces (5589d8b)

• add pipeline, steering hook (d27e7f6)

• add dataset preprocessing (33923b8)

• add datasets (167d8a2)

• initial commit (6899bb2)