docs(lectures): Add concrete examples, code, and explanations to weeks 1-6

Comprehensive enhancements to all slide decks:
- Week 1 (Lectures 1-4): Added Turing Test examples, ELIZA dialogue samples
- Week 2 (Lectures 5-8): Added neural network diagrams, activation functions
- Week 3 (Lectures 9-11): Added transformer architecture, attention examples
- Week 4 (Lectures 12-14): Added prompting techniques, few-shot examples
- Week 5 (Lectures 15-17): Added fine-tuning code, RLHF workflows
- Week 6 (Lectures 18-20): Added evaluation metrics, benchmark examples

All slides now include:
- Concrete code examples (Python snippets)
- Worked mathematical equations
- Real-world application examples
- Visual diagrams where appropriate

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Author: jeremymanning

slides/week1/lecture1.md

@@ -257,6 +257,48 @@ We cannot directly observe consciousness — even in other humans!
 ```
 <!-- caption: Person follows rules but does not understand Chinese -->
 
+---
+<!-- _class: scale-78 -->
+
+# Worked example: The Chinese Room
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+**The scenario:**
+1. You are locked in a room
+2. Chinese characters slide under the door
+3. You look them up in a rule book
+4. You copy out the corresponding response
+5. You slide it back out
+
+**From outside:** Perfect Chinese conversation!
+**From inside:** Just symbol manipulation.
+
+</div>
+<div style="flex: 1;">
+
+**Concrete example:**
+```
+Input:  你好吗? (How are you?)
+Lookup: Rule #42 → 我很好
+Output: 我很好 (I am fine)
+```
+
+You matched symbols without knowing:
+- What "你好吗" means
+- That it's a greeting
+- What "你" (you) refers to
+
+</div>
+</div>
+
+<div class="warning-box" data-title="The key insight">
+
+Syntax (symbol manipulation) is not sufficient for semantics (meaning).
+
+</div>
+
 ---
 
 # Volition
@@ -400,11 +442,48 @@ Having words for things affects how we *see* them:
 
 <div class="note-box" data-title="Key finding">
 
-English speakers are slower to distinguish shades of blue than Russian speakers: 
+English speakers are slower to distinguish shades of blue than Russian speakers:
 language categories affect *perception*, not just description!
 
 </div>
 
+---
+<!-- _class: scale-78 -->
+
+# Worked example: The color experiment
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+**The task:**
+Look at three color squares. Which one is different?
+
+```
+Trial 1: [dark blue] [dark blue] [light blue]
+         Russian speakers: FAST (different words!)
+         English speakers: slower (both "blue")
+
+Trial 2: [blue] [blue] [green]
+         Both groups: equally fast (cross-category)
+```
+
+</div>
+<div style="flex: 1;">
+
+**What this tells us:**
+- Having distinct words for colors creates **perceptual boundaries**
+- The brain literally processes colors faster when they cross linguistic categories
+- Language doesn't just describe reality; it *shapes* how we see it
+
+</div>
+</div>
+
+<div class="tip-box" data-title="Implication for LLMs">
+
+If language shapes perception, what "perception" do LLMs have? They experience language *about* color, but never color itself.
+
+</div>
+
 ---
 
 # So...what about LLMs?
@@ -424,6 +503,78 @@ How do symbols (words) get their meaning?
   - Generate text based on those patterns
   - No direct connection with the external world
 
+---
+<!-- _class: scale-70 -->
+
+# Worked example: Grounding vs. Statistics
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+**Human understanding of "coffee":**
+```
+Sensory memories:
+  - Bitter taste
+  - Rich aroma
+  - Warm ceramic mug
+  - Morning ritual feeling
+  - Caffeine effect on body
+```
+These experiences **ground** the word.
+
+</div>
+<div style="flex: 1;">
+
+**LLM "understanding" of "coffee":**
+```
+Statistical associations:
+  - Often appears with "morning"
+  - Frequently near "cup", "drink"
+  - Associated with "caffeine"
+  - Patterns: "I need my ___"
+  - Context: restaurants, breakfast
+```
+Pure pattern matching, no experience.
+
+</div>
+</div>
+
+<div class="warning-box" data-title="The question">
+
+Can statistical patterns **ever** capture what it *feels like* to drink coffee?
+
+</div>
+
+---
+
+# How LLMs actually work: A preview
+
+<div class="note-box" data-title="Next-token prediction">
+
+At their core, LLMs are trained to predict what comes next. Given a sequence of words, what is the most likely continuation?
+
+</div>
+
+```python
+# Conceptually, this is what an LLM does:
+input_text = "The cat sat on the"
+
+# LLM predicts probability for each possible next word:
+predictions = {
+    "mat": 0.35,      # Most likely
+    "floor": 0.20,
+    "couch": 0.15,
+    "roof": 0.05,
+    "elephant": 0.001  # Very unlikely
+}
+```
+
+<div class="tip-box" data-title="Key insight">
+
+LLMs don't "know" what cats are. They've just seen "cat sat on the mat" many times!
+
+</div>
+
 ---
 
 # Our approach (in this course)

slides/week1/lecture2.md

@@ -188,6 +188,54 @@ new_text = re.sub(r"I am", "You are", text)
 # "You are feeling very happy today"
 ```
 
+---
+<!-- _class: scale-70 -->
+
+# Worked example: Building a simple chatbot rule
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+**Goal:** Respond to "I am [feeling]" statements
+
+```python
+import re
+
+def simple_respond(user_input):
+    # Try to match "I am [something]"
+    match = re.search(
+        r"I am (.*)",
+        user_input,
+        re.IGNORECASE
+    )
+
+    if match:
+        feeling = match.group(1)
+        return f"Why are you {feeling}?"
+
+    return "Tell me more."
+```
+
+</div>
+<div style="flex: 1;">
+
+**Try it:**
+```python
+>>> simple_respond("I am tired")
+"Why are you tired?"
+
+>>> simple_respond("I am feeling anxious")
+"Why are you feeling anxious?"
+
+>>> simple_respond("Hello there")
+"Tell me more."
+```
+
+**Key insight:** The computer has no idea what "tired" or "anxious" mean. It's just copying text!
+
+</div>
+</div>
+
 ---
 
 # Meet ELIZA: a computerized Rogerian therapist
@@ -234,6 +282,54 @@ Have a [conversation with ELIZA](https://contextlab.github.io/llm-course/demos/0
 - What kinds of inputs break the illusion?
 - **How do you think ELIZA *works*?**
 
+---
+<!-- _class: scale-70 -->
+
+# Example ELIZA conversation
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+```
+User:   I am feeling sad today
+ELIZA:  Why are you feeling sad today?
+
+User:   My mother doesn't understand me
+ELIZA:  Tell me more about your family.
+
+User:   She always criticizes everything I do
+ELIZA:  Your mother?
+
+User:   Yes, and my father too
+ELIZA:  What else comes to mind when you
+        think of your father?
+
+User:   I don't know
+ELIZA:  Why don't you know?
+```
+
+</div>
+<div style="flex: 1;">
+
+**Notice how ELIZA:**
+- Picks up on keywords ("mother", "father")
+- Reflects statements back as questions
+- Uses generic prompts when stuck
+- Never makes specific claims about the world
+
+**Where it breaks down:**
+```
+User:   What is 2 + 2?
+ELIZA:  Why do you ask?
+
+User:   I want to know if you're intelligent
+ELIZA:  Does it please you to believe I
+        am intelligent?
+```
+
+</div>
+</div>
+
 ---
 
 # The ELIZA effect
@@ -295,6 +391,48 @@ Have *you* experienced the ELIZA effect with modern AI systems (ChatGPT, Claude,
 - What is the difference between *seeming* intelligent and *being* intelligent?
 - How would we *know* if an AI truly understood us?
 
+---
+<!-- _class: scale-78 -->
+
+# Example: The ELIZA effect in modern AI
+
+<div style="display: flex; gap: 1.5em;">
+<div style="flex: 1;">
+
+**Feels like understanding:**
+```
+User: I'm really stressed about my exam
+AI:   I can hear that you're feeling
+      overwhelmed. Exams can be really
+      stressful. What subject is it?
+
+User: Physics. I've been studying for weeks
+AI:   Studying for weeks shows real
+      dedication. What part of physics
+      is giving you trouble?
+```
+
+**Creates the illusion of empathy**
+
+</div>
+<div style="flex: 1;">
+
+**Reveals the limitation:**
+```
+User: I just realized I left my exam
+      notes in your office
+AI:   I understand that can be frustrating!
+      Would you like tips on how to
+      retrieve your notes?
+```
+
+The AI has no office. It cannot have your notes. But it responds as if this makes sense!
+
+**Pattern matching, not understanding.**
+
+</div>
+</div>
+
 ---
 
 # Reading: Weizenbaum (1966)