Praxis — An adaptive tutor that shows its working Guest learner

AGI Agents (6-week course) › W5 · Reward models, DPO & reward hacking

0/0 this session +0 XP
Numeric Easy bradley-terrysigmoid ~60s

A reward model scores response A at r_A = 1.5 and response B at r_B = 0.5. Under the Bradley–Terry model used in reward_model.py, what is P(A preferred over B)? Give a decimal to 2 dp.

Fractions, decimals and π are all understood.

Hints are free to take and they are counted. A correct answer after two hints moves your mastery less than one without — which is the honest way to score it.

Why this question? this skill is new to you; the difficulty is about right — roughly 72% chance you get it; and it needs calibrating — few people have tried it Predicted chance you get it: 72% — the selector aims at 78%.
difficulty fit 0.815
skill need 0.75
review boost 0
calibration bonus 0.12
recency penalty 0
repetition penalty 0
frustration 0