Montonicity Bias

by Ari Holtzman2 months ago
0

I've noticed that if you give an LLM a list of instructions:

  1. INSTRUCTION 1
  2. INSTRUCTION 2
    ...

it tends to try to follow them in order, even if instruction j relies in on instruction i already having been followed where i < j.

If you are inspired by this idea, you can reach out to the authors for collaboration or cite it:

@misc{holtzman-montonicity-bias-2026,
  author = {Holtzman, Ari},
  title = {Montonicity Bias},
  year = {2026},
  url = {https://hypogenic.ai/ideahub/idea/1CoZIwMuZPRyh8JBlwRK}
}

Comments (0)

Please sign in to comment on this idea.

No comments yet. Be the first to share your thoughts!