pith. sign in

Sahar Abdelnabi

Identifiers

  • name variant Sahar Abdelnabi 0.60 · backfill

Papers (11)

  1. Models That Know How Evaluations Are Designed Score Safer cs.CL · 2026 · author #4
  2. Decomposing and Measuring Evaluation Awareness cs.LG · 2026 · author #5
  3. Measuring Security Without Fooling Ourselves: Why Benchmarking Agents Is Hard cs.CR · 2026 · author #1
  4. AI Agents May Always Fall for Prompt Injections cs.CR · 2026 · author #1
  5. Hidden in Memory: Sleeper Memory Poisoning in LLM Agents cs.CR · 2026 · author #4
  6. No More, No Less: Task Alignment in Terminal Agents cs.LG · 2026 · author #7
  7. Detecting Multi-Agent Collusion Through Multi-Agent Interpretability cs.AI · 2026 · author #3
  8. Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks cs.CR · 2026 · author #3
  9. Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems cs.MA · 2026 · author #4
  10. Safety Must Precede the Deployment of Open-Ended AI cs.AI · 2025 · author #3
  11. Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection cs.CR · 2023 · author #2

Mentions

  • 2502.04512 #3 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2605.28591 #4 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2602.15198 #4 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2605.23055 #5 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2605.22568 #1 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2605.17634 #1 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2605.15338 #4 · arxiv_oai · confidence 0.70 Sahar Abdelnabi
  • 2602.20156 #3 · arxiv_oai · confidence 0.70 Sahar Abdelnabi

Frequent Coauthors