Expired50% confidenceTradeoffExact time
Mainstream large language models show significantly reduced accuracy when understanding obfuscated bundles with extremely long contexts (tens of thousands of lines) and may hallucinate on custom encryption variants.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
Valid until: 10/18/2026(expired)
Sources
Related Claims
UnverifiedCross-lingual jailbreaking exploits the asymmetry between a model's concept space and its language space, as the learning of safety-refusal behavior is highly language-specific.78% similarVerified研究表明大语言模型规模越大，幻觉的伪装性往往越强，能生成更连贯、更具说服力的错误内容76% similarVerifiedCurrent large language models exhibit a pattern in the systems programming domain of being strong on high-frequency knowledge and weak on long-tail knowledge, tending to make reasonable inferences rather than strictly following specifications for edge-case semantics and version-specific compatibility behavior.75% similarVerifiedLarge language models are prone to forgetting early information or hallucinating during long conversations, which in programming manifests as fabricating non-existent APIs or generating logically incorrect code.75% similarUnverifiedLarge language models have a miscalibration problem, often answering in a confident tone regardless of whether the answer is correct or incorrect.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/572686API
curl https://kongchang.com/api/v1/knowledge/claims/572686MCP
get_claim(id=572686)