Unverified50% confidenceBenchmarkExact time
在 LibriSpeech 基准测试上顶尖系统 WER 已低于 2%，在 Switchboard 对话场景中通常为 5-8%，在多人会议场景（AMI、ICSI）中可能高达 15-25%
1
Sources
50%
Confidence
Long-term
Relevance
8/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn chat workloads, pruning low-confidence tokens raises the acceptance rate from 45.7% to 95.7%64% similarUnverified在大型开源AI提示词仓库中，真正被高频使用的往往只占总量的不到5%63% similarUnverified当前主流 ASR 引擎的通用词错率（WER）已降至5%以下，但在医疗垂直领域的专有名词上错误率可能飙升至15%-30%，除非进行领域微调和自定义词典配置61% similarVerifiedThe acceptance rates of top AI conferences (such as NeurIPS, ICML, and ICLR) are typically between 20% and 30%.61% similarUnverifiedWhen a conversation reaches approximately 50% memory capacity, AI begins to degrade in quality, exhibiting repeated ineffective suggestions and declining code quality.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/813397API
curl https://kongchang.com/api/v1/knowledge/claims/813397MCP
get_claim(id=813397)