TL;DR (and please correct me if I got it wrong):
Tiny deterministic model predicts the K/V cache, prediction is compared with reality, delta is stored in vram. The other way round then just predicts the values again, applies the delta, and you have the full correct value while just storing the delta
And this works because you're never looking at the whole k/v cache but always just a slice. So you just need a memory buffer of the size of the slice
___
If this works out and I've understood correctly, that _I think_ would mean that a 24GB RTX 4090 could fit 256k q8 context next to Qwen3.6-27B at IQ4_NL.
Or, alternatively, something like 208k context (matching claude api limits of 200k in some plans) with a slightly larger quant like UD-Q4_K_XL.
That would be massive. Especially since the thing has so much compute to spare.
Though, all depending on the size of that predictor model I guess?
How do these results compare with the engram based approach from deepseek?
[dead]
The problem with this approach is that even recomputing a "draft" of the KV cache is still quadratic in context length. Maybe you can get some constant savings by always recomputing the earliest tokens, but it's not a good tradeoff as context sizes grow.
BTW, I forgot to mention that you can make this work in a way, but only if your model architecture generalizes the context and attention mechanism such that it's no longer a pure sequence. So you could have a large amount of distinct "early" token sequences, with each being self-contained and not depending on any other tokens, e.g. your source code files might be such. Then later parts of the context would of course depend on all of those files as usual. This makes prefill for the earlier context both reusable and cheaply recomputable throughout, at the cost of losing some dependencies that would've been previously accounted for: your model becomes faster and more efficient, but perhaps not quite as smart.
Sure, but any classical attention mechanism is quadratic in context length.
But text generation is quadratic after the KV cache optimization. If every decode step now has to recompute KV cache including its latest and most expensive tokens (even with a quick, "draft" model) that's even worse.
Note that any cache (eg LRU-eviction) is just a specific speculative model for future usage :-)
The cache can be backed by hardware/lookup, or by a cheap computation. The line between functions and data is really blurry.
Would you say it is homoiconic, similar to LISP where the syntax of the language is the AST; so, data can become code (Macros) and code can be data (the S-Expression)?
I am yet to do a "deep dive" into the results, but what a well written article. An LLM could _never_ write so crisply.
There is no compression taking place here.
Isn't that nitpicking? It's a smaller representation of the data, if you have a certain appetite for decompression time. It could conceivably be worth it. I think it would make a great level 2 cache for older chats.
Isn't the delta fed to an arithmetic coder?
It is a “research note”. It might not pan out, and you might say it doesn’t deserve the attention on the internet. But it did suggest something that resembles of compression, just no experiment done for that.
If “speculative” approach works so well in different contexts why not make it first class and use everywhere, possibly recursively?
Speculation is only worth it if you can profit from it. Not every context allows this or has a similar idea of what can be speculated.
It works very well on dense models, imho great alternative to MoE. As verification is cheaper than generation it could be fundamental, first class primitive, maybe even to recurse on it, do live distillation during inference etc.
MoE is more hardcoded, pre determined, speculation is much more dynamic, malleable after training.
This paper actually proposes direction of aligning architecture to aid speculation as future work.
Multi-token prediction is a good enhancement to training. It isn't necessarily useful for inference. Other speculative decoding like EAGLE is. It is specific to the technology and the authors of these things write about it.
You can use the original model to compress the kv cache and get ∞x compression, since the prediction is perfect. The cost is time, and I don't see how this could be worth it.
The tradeoff gets better the bigger your primary model, and probably with bigger batch sizes. The KV cache can consume a lot of expensive VRAM, and the VRAM and compute costs of the predictor model become a small fraction of the cost of the primary model
For serving a 1T model with 16 concurrent requests this could make a lot of sense. For a 8B model with a single request far less so
This can't be used to save VRAM in practice. To generate a new token with the primary model, you first need to decompress the cache, which involves regenerating the whole sequence from scratch. I.e. generate 1 million tokens with the small model to generate 1 with the large.