Unverified60% confidenceOpinionExact time
Large models perform better when processing decompiled pseudocode (such as C-like pseudocode output by Ghidra), because the training corpus contains a large amount of high-level language code, while their ability to understand raw assembly instructions is relatively limited.
2
Sources
60%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified实验数据显示，语法约束解码使小型语言模型生成Bash代码的语法正确率大幅提高，引号不匹配和结构不完整等错误几乎被完全消除79% similarVerifiedLarge language models' performance in code generation is highly dependent on the scale and quality of pre-training data.79% similarVerifiedLarge language models have hallucination defects in code generation scenarios, manifested as referencing non-existent library functions and generating syntactically correct but logically erroneous code.76% similarUnverifiedFor models, verbose and explicit code tends to be easier to accurately generate and parse, because the context is more complete and there is less ambiguity.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/899699API
curl https://kongchang.com/api/v1/knowledge/claims/899699MCP
get_claim(id=899699)