Ala N. Tak
About Me
I’m Ala, a Computer Science Ph.D. candidate at the University of Southern California.
My research combines Large Language Models (LLMs), mechanistic interpretability, safety alignment, model steering, agentic workflows, and human–computer interaction (HCI).
I build LLM-based systems that are emotionally aware, interpretable, and aligned for safe interaction.
News
- Jan 2026 — Submitted to ICML2026: Sparks of Rationality: Do Reasoning LLMs Align with Human Judgment and Choice?.
- Nov 2025 — Submitted to ARR: Psychological Steering in LLMs: Effectiveness & Trustworthiness.
- Aug 2025 — Wrapped up my internship at Visa Research, delivering a scalable agentic RAG platform for enterprise business intelligence.
- Jul 2025 — Poster Presentation: Mechanistic Interpretability of Emotion Inference in LLMs accepted at ACL 2025.
- Jul 2025 — Workshop Organization: Serving as on-site host for the 3rd Workshop on Social Influence in Conversations (SICon) at ACL 2025, Vienna.
- Jul 2025 — Essay Published: How Language Models Form Social Impressions (and What That Reveals) on ICT News.
- Jun 2025 — Paper Accepted: Social Impression Formation in LLMs accepted at CoLM 2025.
- Jun 2025 — Paper Accepted: Aware yet Biased: Investigating Emotional Reasoning and Appraisal Bias in LLMs accepted at IEEE T-AFFC 2025.
- Jun 2025 — Essay Published: Investigating Emotional Reasoning and Appraisal Bias in LLMs on ICT News.
For More Information
More info can be found at ICT.