How many reward-hacking interview questions are there?

1 reported question, mostly Reinforcement Learning.

How hard are reward-hacking questions?

They average 5.0 out of 5: 1 at 5/5.

reward-hacking Interview Questions

1 interview question in our bank covers reward-hacking, most of them Reinforcement Learning. They average 5.0/5 difficulty — very hard — and each one was reported by a candidate after a real interview. Companies known to ask about reward-hacking: Amazon.

Practice these on the problems board →

Companies that ask about reward-hacking

Amazon1

Question mix

Reinforcement Learning1

Difficulty

5/5 — very hard1

Questions tagged reward-hacking

RLHF: PPO vs GRPO vs GSPO
5/5
Dive into advanced reinforcement learning concepts by comparing prominent alignment techniques such as PPO, GRPO, and GSPO, based on applied scientist interviews at Amazon. This deep-dive examines the mathematical formulations behind clipped surrogate objectives, KL divergence penalties, group-based relative advantages, and how recent reasoning-focused models leverage these updates. You will also analyze reward model construction, mitigation strategies for reward hacking, and distributed training topologies. The complete analytical breakdown and model response guide require a subscription.
Reinforcement LearningAmazon

reward-hacking interview FAQ

How many reward-hacking interview questions are there?: 1 reported question, mostly Reinforcement Learning.
Which companies ask reward-hacking questions?: Amazon (1).
How hard are reward-hacking questions?: They average 5.0 out of 5: 1 at 5/5.

Companies that ask about reward-hacking

Question mix

Difficulty

Questions tagged reward-hacking

RLHF: PPO vs GRPO vs GSPO

reward-hacking interview FAQ