Unverified50% confidenceFactExact time
BAAI/bge-small-en-v1.5 uses a WordPiece tokenizer, which splits snake_case and camelCase into meaningless fragments.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe Llama series uses SentencePiece BPE tokenization, while the Qwen series uses a custom tokenizer based on byte-level BPE.56% similarUnverified即使输入完全无意义的假词，文生图模型也不会拒绝处理，而是通过BPE拆解为已知子词并生成看起来连贯的角色图像53% similarUnverifiedHermes v0.20.6's gateway passes transcribed text to the model as a plain quoted string without any voice markers or source information, preventing the model from distinguishing between a casual mention of a memo and a formal command to trigger the voice-memo skill.52% similarUnverifiedSemantic chunking uses embedding models to assess semantic similarity between adjacent sentences and splits at semantic breaks, but still cannot guarantee logical coherence across paragraphs.51% similarUnverifiedPrevious versions could lose or truncate reasoning content when concatenating streaming chunks or processing multi-turn tool call responses.51% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1024149API
curl https://kongchang.com/api/v1/knowledge/claims/1024149MCP
get_claim(id=1024149)