Article
GLM-5.2: Built for Long-Horizon Tasks
zai-org
• • 120
colored_dye
colored-dye
AI & ML interests
mechanistic interpretability, training data attribution, representation steering
Recent Activity
upvoted an article 17 days ago
GLM-5.2: Built for Long-Horizon Tasks upvoted a paper about 2 months ago
PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts upvoted a paper about 2 months ago
Faithful Bi-Directional Model Steering via Distribution Matching and Distributed Interchange InterventionsOrganizations
None yet
upvoted an article 17 days ago
upvoted 3
papers about 2 months ago
PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts
Paper • 2605.05974 • Published • 1
Faithful Bi-Directional Model Steering via Distribution Matching and Distributed Interchange Interventions
Paper • 2602.05234 • Published • 1
Towards Steering without Sacrifice: Principled Training of Steering Vectors for Prompt-only Interventions
Paper • 2605.05983 • Published • 1
upvoted a paper about 1 year ago