CIS 6930: Special Topics in Large Language Models

Course Information

Course Title: CIS 6930 Special Topics in Large Language Models

Semester: Fall 2025

Lectures: Monday Wednesday Friday 11:45am - 12:35pm at CSE E222

Office Hour: Friday 1:00pm - 2:00pm

Course Description

Large Language Models (LLMs) provide the foundational technology behind recent breakthroughs in Natural Language Processing, enabling systems to and perform a wide range of language-driven tasks. This course provides a comprehensive introduction to the fundamentals in Large Language Models, covering their tokenization, model architecture, training methods, inference, fine-tuning, post-training techniques. We will also introduce recent advances in Large Language Models, including retrieval-augmented LLMs, knowledge-enhanced LLMs, reasoning and planning with LLMs, LLMs agent, and multi-modality LLMs. In addition, we will discuss the current limitations of Large Language Models, such as harm, hallucinations, and safety alignment concerns, and will explore potential mitigation strategies for addressing these issues.

Grading

• Quizzes (25%): Tentative quiz dates are Sep 5th, Sep 19th, Oct 3rd, Oct 24th, Nov 7th, Nov 21st 2025

• Mid-term Exam (35%): Oct 10th, 2025 11:45am - 12:35pm at CSE E222

• Final Projects (40%): Project Proposal (10%) due Oct 5th 2025, In-class Presentation (15%), Final Project (15%) due Dec 5th 2025

Course Schedule (tentative)

Week #                                           Topics

Week 1                                                  Lectures Introduction

                                                               LLMs capabilities and limitations

Week 2                                                  N-gram Language Model

                                                               LLMs Tokenization Algorithms, Byte-Pair Encoding, WordPiece, Unigram Tokenizer

Week 3                                                  Neural Architectures for Language Model, Feedforward Neural Language Model

                                                               Recurrent Neural Network for Language Model, Long Short-Term Memory Model

Week 4                                                  Sequence-to-Sequence Language Model

                                                               Attention Mechanism, Self-Attention, Multi-head Attention

                                                               Transformer Language Model

Week 5                                                  Encoder Language Model, Encoder-Decoder Language Model, Decoder-only Language Model

                                                               Pre-Training of Large Language Models, Data Curation for LLMs

Week 6                                                  Large Language Model Scaling Law and Emergent Behavior

                                                               Inference of Large Language Models, LLMs Decoding Strategies, Controllable Text Generation

Week 7                                                  Fine-Tuning of Large Language Models

                                                               Parameter-Efficient Fine-Tuning of Large Language Models

Week 8                                                  Post-Training of Large Language Models, In-context Learning, Instruction Tuning, Alignment

                                                               Reinforcement Learning in Large Language Models, Proximal Policy Optimization (PPO), 

                                                               Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO)

Week 9                                                  Retrieval-augmented LLMs, Retrieval Augmentation Generation

                                                                Knowledge-enhanced LLMs, Knowledge Boundary, Knowledge Steering, Knowledge Editing etc.

Week 10                                                Reasoning and Planning in Large Language Models

                                                                LLMs Agent and Multi-Agent systems

Week 11                                                LLMs Hallucination Detection and Mitigation

                                                                LLMs Harm, Debias, Detoxification

Week 12                                                Large Multi-Modality Models, Vision-Language Foundation Models

Week 13                                               Students Projects Presentation

Week 14                                               Students Projects Presentation