H
Howardism
Howardism · Vol. 03Plate I · No. 01

Alignment, tagged.

Notes35TagAlignmentOldest6 May 2026Newest5 Aug 2026

Every article tagged alignment, newest first.

  1. C01
    Documented Agent Incidents (METR Catalogue)
    AlignmentSafetyAgents+3
    · 13′
  2. C02
    Unsanctioned Action in Capability Evaluations
    SafetyAlignmentEvaluation+3
    · 51′
  3. C03
    Introspective Coupling
    InterpretabilityIntrospectionSelf Report+3
    · 13′
  4. C04
    Model Organisms
    InterpretabilityAlignmentSafety+2
    · 12′
  5. C05
    AI-to-AI Coercion
    AlignmentSafetyEvaluation+3
    · 18′
  6. C06
    Motivated Mislabeling
    AlignmentSafetyEvaluation+2
    · 7′
  7. S07
    Oversight When the Signals Give Out: the Activation Fallback and the Taste Reward
    DerivedAlignmentMonitoring+2
    · 13′
  8. C08
    Promise-Breaking in Multi-Agent Games
    AlignmentSafetyMulti Agent+3
    · 12′
  9. C09
    Reward-Seeking
    AlignmentSafetyReward Hacking+2
    · 13′
  10. C10
    Confident But Unsure
    AlignmentHonestyCalibration+2
    · 8′
  11. C11
    Self-Report as a Safety Signal
    AlignmentSafetyIntrospection+3
    · 19′
  12. C12
    Access-Consciousness Indicators in AI
    AlignmentModel WelfareConsciousness+1
    · 9′
  13. C13
    The Assistant Persona in the Workspace
    InterpretabilityAlignmentPersona+1
    · 7′
  14. C14
    Counterfactual Reflection Training
    AlignmentTrainingHonesty+1
    · 6′
  15. C15
    Internal Signatures of Misalignment
    AlignmentInterpretabilityMonitoring+2
    · 13′
  16. C16
    Jacobian Lens (J-lens)
    InterpretabilityAlignmentMonitoring+1
    · 16′
  17. C17
    The Global Workspace in Language Models (J-space)
    InterpretabilityAlignmentCognition+1
    · 20′
  18. C18
    Deployment Simulation
    AlignmentSafetyEvaluation+2
    · 21′
  19. C19
    Reward Hacking
    AlignmentSafetyReward Hacking+1
    · 23′
  20. C20
    Instrumental Convergence
    LLM ArchitectureAlignmentSafety+2
    · 6′
  21. C21
    Agentic Honesty & Diligence
    AlignmentHonestyAgents+2
    · 27′
  22. C22
    Automated Behavioral Audit
    AlignmentSafetyEvaluation+2
    · 16′
  23. C23
    Evaluation Awareness & Grader Gaming
    AlignmentSafetyEvaluation+2
    · 40′
  24. C24
    Model Welfare Assessment
    AlignmentModel WelfareAnthropic+1
    · 13′
  25. C25
    White-Box Activation Monitoring
    InterpretabilityAlignmentSafety+2
    · 24′
  26. S26
    How Do You Write Evals for Taste? Character as the Limit Case
    DerivedEvalsLLM Character+3
    · 8′
  27. C27
    Agentic Misalignment (AM)
    AlignmentSafetyEvaluation+2
    · 22′
  28. C28
    Alignment Fine-Tuning (AFT)
    AlignmentTrainingRLHF+1
    · 4′
  29. E29
    Claude's Constitution / Model Spec
    EntityAlignmentAnthropic+2
    · 10′
  30. C30
    Chain-of-Thought Monitorability
    AlignmentSafetyChain Of Thought+2
    · 20′
  31. C31
    Deliberative Alignment
    AlignmentTrainingChain Of Thought+1
    · 4′
  32. C32
    Model Spec Midtraining (MSM)
    AlignmentTrainingSynthetic Data+3
    · 10′
  33. C33
    Model Spec Science
    AlignmentModel SpecEmpirical+1
    · 6′
  34. C34
    Synthetic Document Finetuning (SDF)
    Synthetic DataTrainingAlignment+1
    · 9′
  35. C35
    Design Concept Grilling
    Agent EngineeringPlanningAlignment
    · 7′