arxiv4d ago
arXiv:2607.20469v1 Announce Type: new Abstract: Large language models (LLMs) handle many tasks with one set of parameters, but under KV-cached inference it is unclear what task-general structure, if any, is used at decode time rather than during prefill. We propose DecodeShare, a protocol that ident
arxiv6d agobullish
arXiv:2607.18253v1 Announce Type: new Abstract: Modern language query routers improve inference efficiency by assigning each query to a model that balances response quality and monetary cost. However, current query routers are largely latency-agnostic and do not consider the generation latency exper
arxivJul 20
arXiv:2607.15870v1 Announce Type: new Abstract: Human label variation in natural language inference is increasingly treated as signal rather than noise, but how much of it formal semantic structure explains has not been measured directly. We measure it on the 3,113 SNLI and MNLI items of ChaosNLI, u
arxivJul 16bullish
arXiv:2505.18231v3 Announce Type: replace-cross Abstract: Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache. Vector Quantization (VQ) is recently adopted to alleviate this
arxivJun 10bullish
arXiv:2606.10820v1 Announce Type: cross Abstract: Autoregressive (AR) language modeling is the dominant paradigm for text generation, yet its sequential token-by-token decoding makes inference memory-bound and inefficient. Existing acceleration approaches, such as speculative decoding and diffusion
arxivJun 5bullish
arXiv:2606.05557v1 Announce Type: new Abstract: A situated query like "where is Lin Wei?" often encodes more than its literal content: the user may also want to know whether Lin Wei is free, in a good mood, or worth interrupting now. Standard tool-use agents answer the literal question and stop. AUR
arxivJun 2bullish
arXiv:2606.00618v1 Announce Type: new Abstract: Generative models have emerged as a powerful paradigm for AI planning, yet their performance remains constrained by the training data distribution. One approach is to improve generated solutions during inference by scaling test-time compute. A more eff
arxivMay 14bullish
arXiv:2605.13784v1 Announce Type: new Abstract: Conventional transformer inference engines are request-driven, paying an O(n) prefill cost on every query. In streaming workloads, where data arrives continuously and queries probe an ever-growing context, this cost is prohibitive. We introduce a data-
arxivMay 8bullish
arXiv:2605.05225v1 Announce Type: cross Abstract: Mixture-of-Experts Multimodal Large Language Models (MoE MLLMs) suffer from a significant efficiency bottleneck during Expert Parallelism (EP) inference due to the straggler effect. This issue is worsened in the multimodal context, as existing token-
arxivMay 6
arXiv:2605.00955v1 Announce Type: cross Abstract: Retrieval-Augmented Generation (RAG) equips large language models (LLMs) with external evidence by retrieving documents at inference time, but it also turns the retrieval corpusinto a sensitive asset. Under a black-box setting, an adversary given a c
arxivApr 22bullish
arXiv:2603.16091v2 Announce Type: replace-cross Abstract: In factual question answering, many errors are not failures of access but failures of commitment: the system retrieves relevant evidence, yet still settles on the wrong answer. We present CounterRefine, a lightweight inference-time repair lay
arxivApr 10bullish
arXiv:2604.06779v1 Announce Type: new Abstract: We introduce Fleming-Viot Diffusion (FVD), an inference-time alignment method that resolves the diversity collapse commonly observed in Sequential Monte Carlo (SMC) based diffusion samplers. Existing SMC-based diffusion samplers often rely on multinomi