Unverified50% confidenceBenchmarkExact time
Transferring the PLE (n-gram memory) pre-trained by Qwen3.8 Flash-Next to the Qwen3.5-0.8B model achieved a 5.05% perplexity reduction on the frozen validation set.
1
Sources
50%
Confidence
Long-term
Relevance
9/25/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the experiment, the Qwen3.5-0.8B backbone and approximately 51 billion parameter PLE memory module were frozen, with only R=1 readers deployed at decoder layers 3 and 9 being trained.78% similarUnverifiedQwen3.8 Flash Next is equipped with an independent Ngram embedding system that adds approximately 51 billion additional parameters; these parameters do not participate in regular forward propagation and are activated only upon specific pattern matches71% similarUnverifiedThe benchmark results of Qwen3.8 Flash Next are official Qwen data and have yet to be independently verified by third parties.70% similarUnverifiedQwen3.8-Flash-Next natively supports MTP, indicating that this capability was built in during the model training phase.69% similarUnverifiedA Qwen Flash model released by Alibaba at the end of August also places 51 billion n-gram parameters in system memory for execution.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/949983API
curl https://kongchang.com/api/v1/knowledge/claims/949983MCP
get_claim(id=949983)