Large Language Models and Knowledge-aware AI (Seminar SS2026)

• The number of participants is limited, with priority given to Master students
• To express interest, send an email to ..., including a short motivation statement and your transcript
• Places will be allocated based on background match (courses taken) and motivation

Topics

• LLM versus human language learning
  BabyLM challenge
  Charpentier et al.
• LLM biases
  Bias runs deep: Implicit reasoning biases in persona-assigned llms

  S Gupta, V Shrivastava, A Deshpande, A Kalyan
  Investigating subtler biases in llms: Ageism, beauty, institutional, and nationality bias in generative models
  M Kamruzzaman, MMI Shovon, GL Kim

• Historical perspective: Memex
  As We May Think
  Vannevar Bush
• HALoGEN: Fantastic LLM Hallucinations and Where to Find Them
• FoodTaxo: Generating Food Taxonomies with Large Language Models
• Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet
• Extract, define, canonicalize: An llm-based framework for knowledge graph construction
• Mission: Impossible Language Models
• Unexpected Knowledge: Auditing Wikipedia and Grokipedia Search Recommendations
• Evaluating the Ripple Effects of Knowledge Editing in Language Models
• WikiBigEdit: Understanding the Limits of Lifelong Knowledge Editing in LLMs
• Unveiling Factual Recall Behaviors of Large Language Models through Knowledge Neurons
  

  (own topic suggestions are welcome as well)

Deliverables

There are 5 deliverables. To pass the course, all have to be submitted on time. Percentages in brackets denote contribution to final grade.

1. Outline of report (5%)
2. Report 1st version (0%)*
3. Reviews on two other reports (15%)
4. Reports final version (40%)
5. Presentation (40%)

• 1st revision is not graded, but the prime chance to obtain feedback from advisor and peers.

Tentative timeline

• TBD

Material

• Slides 1st meeting
• Slides 2nd meetings
• Report template