Unverified50% confidenceBenchmarkExact time
The combination of Pythia's first 5 layers plus RWKV's remaining 27 layers reduced the Transformer's KV cache by 84.4%, with no significant difference in accuracy compared to using RWKV alone.
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
Unverified将FP32权重降至INT8后，典型Transformer模型基准测试分数下降通常不超过1%，推理速度提升2至4倍，内存占用减少75%75% similarUnverified批大小为1的Transformer推理算术强度往往低于10 FLOP/Byte，意味着GPU算力利用率不足理论峰值的5%67% similarUnverifiedConformer achieved approximately 15% relative error rate reduction on the LibriSpeech benchmark compared to a pure Transformer architecture.64% similarUnverifiedMLA compresses the KV cache per token in GLM from 32,768 numbers down to 576 numbers, a reduction of approximately 98%.64% similarUnverified基于Transformer架构的现代审核模型相比传统关键词过滤，误报率（False Positive Rate）可降低60%以上64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/964130API
curl https://kongchang.com/api/v1/knowledge/claims/964130MCP
get_claim(id=964130)