Trustworthy ML Course (Spring 2025): Lecture Slides

Lecture 1. Syllabus & Course Overview

Lecture 2. N-Grams, RNNs, LSTMs, Seq2Seq

Lecture 3. Attention Mechanism & Transformers

Lecture 4A. BERT (Bidirectional Encoder Representations from Transformers)

Lecture 4B. GPT-2 (Language Models are Unsupervised Multitask Learners)

Lecture 5A. Saling Laws (Scaling Laws for Neural Language Models)

Lecture 5B. GPT-3 (Language Models are Few shot Learners)

Lecture 6A. FLAN (Fine-tuning Language models are Zero-shot learners)

Lecture 6B. RLHF (Training language models to follow instructions with human feedback)

Lecture 7A. CODEX (Evaluating LLMs Trained on Code)

Lecture 7B. Chain-of-Thought (elicits reasoning in LLMs)

Lecture 8A. LLaMa (Open and efficient foundation language models)

Lecture 8B. Retrieval-Augmented-Generation (RAG)

Lecture 9A. Tree-of-Thought (Evaluating LLMs Trained on Code)

Lecture 9B. DeepSeek-R1 Paper

Lecture 10. Jailbreak attacks (How does LLM safety training fail?)

Lecture 11A. Greedy Coordinate Gradient (GCG) Jailbreak

Lecture 11B. PAIR (Prompt Automatic Iterative Refinement) Jailbreak