fix: Improve page generation and upgrade GPT bot models

- Rewrite build-pages.py with proper markdown parsing:
  - Tables converted to HTML <table> elements
  - Ordered lists converted to <ol><li> elements
  - LaTeX \href commands converted to links
  - Internal links no longer open in new tabs

- Restructure syllabus.md with table format:
  - Course info table at top
  - Weekly overview schedule table
  - Detailed day-by-day tables per week

- Replace crashing Qwen3 with robust fallback chain:
  - Primary: DeepSeek-R1-Distill-Qwen-1.5B (reasoning model)
  - Fallback 1: Gemma 3 1B IT
  - Fallback 2: Gemma 3 270M IT
  - Updated all UI references and architecture tab
  - Added HuggingFace model card links

- Add test-gpt-bot-loading.mjs for model config verification

Author: jeremymanning

admin/syllabus.md

@@ -7,6 +7,7 @@ Welcome to the assignments for PSYC 51.17: Models of Language and Communication.
 All assignments are submitted via GitHub Classroom. Click the "Accept Assignment" button on each assignment page to get started. This will create a personal repository where you'll complete your work.
 
 **Submission Process:**
+
 1. Click "Accept Assignment" to create your repository
 2. Clone the repository to your local machine or open in Google Colab
 3. Complete the assignment following the instructions
@@ -31,6 +32,7 @@ Assignments receive a 10% deduction for each week late, rounded up to the neares
 ## Grading
 
 Each assignment is graded on:
+
 - **Correctness**: Does your code produce the expected outputs?
 - **Code Quality**: Is your code well-organized and documented?
 - **Understanding**: Do your explanations demonstrate understanding of the concepts?
@@ -46,19 +48,25 @@ Each assignment is graded on:
 ## Assignments
 
 ### [Assignment 1: ELIZA Chatbot](./assignment-1/)
+
 Build a pattern-matching chatbot based on Weizenbaum's classic ELIZA program. Learn about string manipulation, regular expressions, and the foundations of conversational AI.
 
 ### [Assignment 2: SPAM Classifier](./assignment-2/)
+
 Develop a text classification system to identify spam messages. Explore feature engineering, tokenization, and evaluation metrics.
 
 ### [Assignment 3: Wikipedia Embeddings](./assignment-3/)
+
 Compare different text embedding methods on Wikipedia articles. Visualize semantic relationships and evaluate embedding quality.
 
 ### [Assignment 4: Customer Service Chatbot](./assignment-4/)
+
 Create a context-aware customer service chatbot using transformer-based models. Implement retrieval and response generation.
 
 ### [Assignment 5: Build GPT](./assignment-5/)
+
 Implement and train a small GPT model from scratch. Understand the transformer architecture, attention mechanisms, and autoregressive generation.
 
 ### [Final Project: Research Project](./final-project/)
+
 Conduct an independent research project applying concepts from the course. Present your findings to the class.