Unverified50% confidenceFactExact time
LEACE fitted on English removed less than one percentage point of Bengali metaphor category accuracy, whereas LEACE fitted on Bengali removed 4-11 percentage points.
1
Sources
50%
Confidence
Long-term
Relevance
11/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder naive byte-level tokenization, Bengali is compressed to an average of approximately 1.4 characters per token, compared to approximately 3.9 for English, causing the linguistic balance of the corpus to be inverted.76% similarUnverifiedTransferring from a distant language such as English to Bengali coreference resolution degrades performance.75% similarUnverifiedIn Bengali, the main verb is often followed by a vector verb, which loses its original meaning and only adds grammatical nuances such as completion.68% similarUnverifiedIndian languages account for a far smaller share of the training corpora of large language models than English, which leads to relatively weak safety guardrails for the model on languages such as Hindi.62% similarUnverified孟加拉语-英语的高质量平行数据仅有数百万级别,而英法语对拥有数千万句对齐语料62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1153331API
curl https://kongchang.com/api/v1/knowledge/claims/1153331MCP
get_claim(id=1153331)