Diffusion language models can write several tokens in one forward pass, and they decide during generation how many passes to spend. Two outputs of the same token length can therefore vary a lot in cost. To understand the underlying factors of this cost variation, we analyze two diffusion models across 10 tasks. We find...
Mohsen Fayyaz, Ali Modarressi, Daniel Mingyi Israel et al.· 0 citations
This work argues that future entropy is also essential for LLM decoding, and estimates its lower and upper bounds by computing the lower and upper bounds on a hidden Markov model (HMM), which is distilled as a surrogate of the LLM.
Zhizhen Chen, Daniel Mingyi Israel, Guy Van den Broeck et al.· 0 citations
This paper formalizes the problem of KV eviction and proves that it is computationally hard, and shows that this probabilistic version of KV eviction coupled with decode time correction is more robust to different tasks compared to existing eviction methods and achieves competitive performance at the same compression b...
Renato Lui Geh, Alexander K. Chen, Daniel Mingyi Israel et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.