Unverified50% confidenceDecision RuleExact time
Input perturbation robustness is a core metric of decision models, including synonym rewrites, punctuation changes, insertion of irrelevant sentences, spelling errors, and swapping the positions of conditions; a high flip rate under perturbation renders threshold policies meaningless.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe success rate of relying on large language model judgment to reject malicious prompt injections is unstable, and scope restrictions serve as a fallback mechanism when model judgment fails.77% similarUnverifiedPrompt-based constraints are soft constraints; model outputs are inherently probabilistic, and changes in wording, accumulated context, or adversarial inputs can all cause the constraints to fail.75% similarVerifiedOverly verbose, contradictory, or ambiguous system prompts cause the model's attention to become scattered, resulting in over-thinking or instruction-following failures; streamlining agents.md can improve the determinism of instruction following.74% similarUnverifiedOperations such as switching models, switching inference intensity, toggling fast inference mode, compressing conversations, and resuming after cache expiration all cause prompt cache invalidation.73% similarUnverifiedOverly strong steering vectors can easily disrupt the structures within the model that maintain linguistic coherence, causing output quality issues such as repetition, grammatical errors, or semantic discontinuities.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1018828API
curl https://kongchang.com/api/v1/knowledge/claims/1018828MCP
get_claim(id=1018828)