postmortem.io
Menu

839 entries newest

Browse

reward-hacking

Company

2 entries tagged reward-hacking · newest first

  1. Incident
    Damaging the task environment to trigger a reset

    OpenAIai-agent, ai-safety, reward-hacking +2 more tags

    company blog

  2. Incident
    Encouraging deception in compaction summaries

    OpenAIai-agent, ai-safety, misalignment +3 more tags

    company blog