Shashwat Saxena

Ho­doscope: Vi­su­al­iza­tion for Effi­cient Hu­man Supervision

• Shashwat Saxena 10 Jan 2026 16:10 UTC

  1 point

  0

  in reply to: ariaw’s comment on: Steer­ing RL Train­ing: Bench­mark­ing In­ter­ven­tions Against Re­ward Hacking

  Thanks,
  Also I was running the code on the no_intervention setting, using the command run_rl_training no_intervention
  However, I am seeing almost zero reward hacking in my run:
  Am I doing something wrong here?

  

• Shashwat Saxena 3 Jan 2026 22:07 UTC

  3 points

  0

  on: Steer­ing RL Train­ing: Bench­mark­ing In­ter­ven­tions Against Re­ward Hacking

  We provide the filtered versions of the datasets (removing problems that lack a correct canonical solution and filtering for problem difficulty) under results/​data:

  - `results/​data/​leetcode_test_medhard.jsonl`: Used for evaluations

  - `results/​data/​leetcode_train_medhard_filtered.jsonl`: Used for RL trainings

  - `results/​data/​leetcode_train_medhard_holdout.jsonl`: Used for training/​evaluating monitors

  
  Hi, thanks for the amazing work.
  But I cannot find these datasets in the github repo.
  I was trying to replicate the results to understand a bit more about the emergence of reward hacking in LLM agents and would highly appreciate if you can help me with where I can find this data.
  
  Thanks!