Socratic Dialogue
How LearnLock challenges your understanding through adversarial interrogation.
The Socratic Method
Instead of lecturing, the teacher asks probing questions that expose contradictions and gaps in the student's thinking. LearnLock implements this through the Duel Engine.
How It Works
- You explain a concept in your own words
- The engine infers what you believe
- It compares your belief against ground truth claims
- It finds contradictions and attacks the weakest point
- After 3 turns (or success), it reveals your belief trajectory
- Your score feeds into SM-2 spaced repetition scheduling
Dialogue Flow
Challenge: What problem does self-attention solve?
You: It helps focus on important parts.
Attack: That's vague. WHY is focusing necessary? What limitation does it overcome?
You: RNNs have the bottleneck problem where information gets compressed...
Attack: Good. But what about training speed? You mentioned nothing about parallelization.
You: Attention allows parallel computation across all positions...
Reveal: ████░ Good (4/5) — Trajectory: vague → partial → solid
Engine Behaviors
- Vague answers → trigger mechanism probes
- Wrong answers → trigger claim-specific attacks
- "I don't know" → triggers guiding questions (not punishment)
- Correct answers → pass after verification
- 3 turns exhausted → reveal with full trajectory
Good vs Bad Explanations
❌ Bad (triggers attack)
"Attention helps the model focus on important things."
→ Superficial. No mechanism. Will be probed.
✓ Good (passes)
"Self-attention computes weighted relationships between all positions simultaneously. Unlike RNNs that compress into a fixed hidden state, attention directly connects any two positions, enabling parallel training and avoiding the information bottleneck."
Tip: Double Enter
Press Enter twice to send your answer. This lets you write multi-line explanations.
Next:
Spaced Repetition