University of Birmingham

university

http://www.bham.ac.uk

AI & ML interests

None defined yet.

Recent Activity

rzdiversity submitted a paper 8 days ago

When Does Correction Become Repair? Mechanistic Auditing of Internal Interventions in Tool-Using LLMs

rzdiversity submitted a paper 10 days ago

See it, Say it, Sorted: Mechanistic Diagnosis and Parameter-Space Mitigation of Emergent Misalignment in LLMs

View all activity

Papers

See it, Say it, Sorted: Mechanistic Diagnosis and Parameter-Space Mitigation of Emergent Misalignment in LLMs

When Does Correction Become Repair? Mechanistic Auditing of Internal Interventions in Tool-Using LLMs

View all Papers

UniversityofBirmingham 's papers 3

Submitted by

Ruizhe Li

See it, Say it, Sorted: Mechanistic Diagnosis and Parameter-Space Mitigation of Emergent Misalignment in LLMs

UniversityofBirmingham

University of Birmingham

Submitted by

Ruizhe Li

When Does Correction Become Repair? Mechanistic Auditing of Internal Interventions in Tool-Using LLMs

UniversityofBirmingham

University of Birmingham

Submitted by

Ruizhe Li

Locked at the Entrance, Open Inside: Where RLVR Narrows the Solution Space

UniversityofBirmingham

University of Birmingham