Back to home

Compare

Comparing: Qwen's Engram Makes Small Models Smarter—Local Trillion-Param Runs Are Fantasy & 通义千问新架构让小模型变聪明 — 但别以为你能本地跑万亿参数

AEN
QwenEngramLocal AI·

Qwen's Engram Makes Small Models Smarter—Local Trillion-Param Runs Are Fantasy

What this is

A 51-billion-parameter "dictionary" hangs on the model, with only ~6 billion parameters activated per inference—that's the core trick of Qwen's newly published Engram architecture: offload rote memorization to lookup tables, freeing up neural capacity for actual reasoning.

The mechanism: build a large table storing vectors (numerical semantic representations) for common phrases like "New York" or "import std". At inference, look up the last 2-3 tokens (the model's minimum text unit) in the table, retrieve results in O(1) time, and feed them to the network—no compute burned.

Small models no longer waste several layers repeatedly spelling out "Wall Street"—they can devote that depth to more complex tasks.

Industry view

Supporters call this one of the most important architectural innovations in the past two years: parameter count no longer equals the ceiling on intelligence, and the barrier to local deployment (running models on your own PC or server) drops significantly.

But Reddit and the technical community poured cold water: first, the memory table is "dumb"—it only looks at the last 2-3 tokens, so a 200K-token context won't change what it retrieves; second, the larger N gets, the sparser the signal—above 4-gram, the paper's own ablation studies (tests that disable components one by one to measure impact) reject the approach; third, memory doesn't substitute for C++ experience—reasoning capability is a separate matter.

Posts shouting about "running 1T models locally" fundamentally misunderstand the mechanism.

Impact on regular people

Enterprise IT: Deploying a "good enough" AI no longer requires chasing the largest parameters—hardware budgets can come down.

Individual professionals: The odds of running near-flagship model performance locally on a laptop are rising—sensitive data doesn't need to hit the cloud.

Consumer market: On-device AI assistants will respond faster and more accurately; expect a new wave of experience upgrades in edge AI apps.

BZH
Qwen通义千问Engram·

通义千问新架构让小模型变聪明 — 但别以为你能本地跑万亿参数

这是什么

510 亿参数当"字典"挂着,每次推理只激活约 60 亿——这是 Qwen 团队新公开的 Engram 架构的核心戏法:让模型把"死记硬背"外包给查表,把神经元腾出来干真正的推理。

具体做法是建一张大表,把"纽约""import std"这种常见词组对应的向量(数值化的语义表示)直接存进去。推理时按最后 2-3 个 token(模型处理文本的最小单位)查表,O(1) 拿到结果喂给网络,不消耗算力。

小模型因此不再花几层网络反复拼写"华尔街"三个字,能腾出深度做更复杂的事。

行业怎么看

支持者认为这是近两年最重要的架构创新之一:参数规模不再等于智能上限,本地部署(在自己电脑或服务器上跑模型)的门槛显著降低。

但 Reddit 和技术社区也泼了几盆冷水:第一,记忆表是"傻"的,只看最后 2-3 个 token,20 万字上下文改变不了它查到什么;第二,N 越大信号越稀,4-gram 以上方案论文自己的消融实验(逐一关停组件看效果的测试)就否了;第三,记忆补不出 C++ 经验,推理能力是另一回事。

那些喊"1T 模型能本地跑"的帖子,对这套机制存在根本误解。

对普通人的影响

企业 IT:部署一个"够用"的 AI,不必再追最大参数,硬件预算有望下台阶。

个人职场:笔记本本地跑接近大模型表现的版本几率上升,敏感数据不必上云。

消费市场:手机端 AI 助手反应更快、回答更准,端侧 AI 应用会有新一波体验升级。