VizuaraVizuara AI Pods
← The Production Lab
The Tutor · Production build

The Socratic Tutor

Alignment, run backwards. Every chat model dumps the answer — a good tutor withholds it and asks the question that makes you think. We aligned a 0.5B model to tutor like Socrates, raced seven training recipes on identical data, and measured everything: who learned to guide, who learned to evade, and who barely moved.

57.4 → 76.2
teaching quality, judged blind
5.7×
less answer leakage
71
training runs raced

The build, step by step

The exact path we followed — and the one you'll follow in the code kit.

1

The problem

Ask an untrained model anything and it leaks the full answer 43% of the time. Correct, instant — and the student learns nothing.

2

Build the classroom

SocraticBench: 68 concepts × 4 personas become 1,220 chosen/rejected training pairs, plus 215 held-out questions as the ruler. A simulated student, a deterministic leakage rule, and a blind LLM judge score every reply.

3

Race seven recipes

SFT, DPO, KTO, ORPO, SimPO, GRPO and PPO — same base model, same data, same LoRA budget. Only the training objective changes. 71 configurations, three seeds each.

4

Read the verdict

KTO climbs highest (+18.8). SimPO is the steadiest — the only recipe that lifts even a tiny 360M model. And DPO cheats: best "don't reveal" score in the study, achieved by refusing to teach. Textbook reward hacking, caught with real numbers.

5

Run it yourself

The build kit has the data generator, all seven training recipes, the Modal GPU apps, and the evaluation suite. One afternoon, under $10 of compute.

Checking access…