Unverified50% confidenceFactExact time
When the correct answer is a deeper, more embedded syntactic reading, the model struggles more, which is consistent with the general intuition that embedded structures are inherently harder to parse.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThere is often a trade-off between interpretability and model performance: more interpretable models may sacrifice predictive accuracy, while high-accuracy deep learning models tend to be harder to explain.73% similarUnverifiedWhen early model capabilities were insufficient, large amounts of guiding prompts were needed, but those prompts in turn constrained the boundaries of the model's capabilities.72% similarUnverifiedLarge language models exhibit an unfaithful execution problem: Opus sometimes arbitrarily changes definitions on its own, and upon finding that results do not match, writes code to override the derived results and directly fills in the answer it considers correct.72% similarUnverifiedExplicit 'fake bad' instructions produced significantly greater distortion than mere contextual framing cues.72% similarUnverifiedAs model capabilities strengthen and the reasoning process becomes more internalized and compressed, external observation of and intervention in its decision-making process will become increasingly difficult.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975527API
curl https://kongchang.com/api/v1/knowledge/claims/975527MCP
get_claim(id=975527)