← People

Ayesha Imran

Source-listed: Software Engineer - AI · ChipHub

  • Language models
  • Robotics
  • Interpretability

Selected work

5
  1. Persona-Vector Routing: A Lightweight, Interpretable Guardrail for Mitigating LLM HallucinationsContributor · 2025
  2. Safety Compass: Monitoring Safety-Relevant Concept Directions During LLM Fine-TuningContributor · 2026
  3. Inoculate or Reflect: Gating vs. Overwriting in Sycophancy ControlMaintainer · 2026
  4. Hallucination Vector ProjectContributor
  5. Confidential CopilotContributor · 2025