concept
active
concept:pre-fill-attackPre-fill attack
Putting words in the target model's mouth to elicit behavior.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Production audit.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Adversarial multi-turn experiment where first turn uses pre-finetuning model to test if follow-up maintains character
- The process by which a model reconstructs the KV cache from the transcript when moving to a new server or different model
- Security attack that bypasses LLM safety alignment by suppressing deliberation or exploiting reflection inhibition.
- Task where a random word is prefilled as the assistant's response, then the model is asked whether it intended to say that word, testing introspection on prior intentions.
- Hidden triggers that elicit malicious behavior.
- Investigator never won using pre-fill attacks and rarely with raw query tool, unlike human teamsfinding0.708Agent fails to use tools like humans.
- Key claim enabling the virtual instance view to survive server changes
- Optimization-based jailbreak method appending strings to prompts to elicit harmful outputs.