Our weekly reading group where we discuss recent papers about LLMs.

Presented by Ashish Kumar on 09/06/2025

Readings:

• ProTeGi
• EvoPrompt

Find the presentation here

Find the recording here

Presented by Ashish Kumar on 09/13/2025

Readings:

• LLMs are Bayesian, In Expectation, Not in Realization

Find the recording here

Presented by Berk Atil on 09/20/2025

Readings:

• Safer or Luckier? LLMs as Safety Evaluators Are Not Robust to Artifacts
• Is LLM-as-a-Judge Robust? Investigating Universal Adversarial Attacks on Zero-shot LLM Assessment

Find the presentation here

Find the recording here