相关论文: MarginGate: Sparse Margin-Triggered Verification f…

General-purpose LLM judges capable of human-level evaluation provide not only a scalable and accurate way of evaluating instruction-following LLMs but also new avenues for supervising and improving their performance. One promising way of…

Modern language models operate on subword-tokenized text in order to make a trade-off between model size, inference speed, and vocabulary coverage. A side effect of this is that, during inference, models are evaluated by measuring the…

When Large Language Models produce structured outputs such as travel plans, code solutions, or multi-step proofs, individual reasoning steps may appear correct while the output as a whole violates budgets, fails test cases, or contradicts…

Large Reasoning Models (LRMs) allocate substantial inference-time compute to Chain-of-Thought (CoT) reasoning, improving performance on mathematics, scientific QA, and tool usage. However, this introduces overthinking: LRMs often reach a…

Long-context inference in large language models is bottlenecked by Key--Value (KV) cache loading during the decoding stage, where the sequential nature of generation requires repeatedly transferring the KV cache from off-chip High-Bandwidth…

flip is an extremely simple and maximally local classical decoder which has been used to great effect in certain classes of classical codes. When applied to quantum codes there exist constant-weight errors (such as half of a stabiliser)…

Large language models (LLMs) excel in many natural language tasks, yet they struggle with complex mathemat-ical problem-solving, particularly in symbolic reasoning and maintaining consistent output. This study evalu-ates 10 LLMs with 7 to 8…

Generative Artificial Intelligence models, such as Large Language Models (LLMs) and Large Vision Models (VLMs), exhibit state-of-the-art performance but remain vulnerable to hardware-based threats, specifically bit-flip attacks (BFAs).…

Here we present a technique for using machine learning (ML) for single-qubit gate synthesis on field programmable logic for a superconducting transmon-based quantum computer based on simulated studies. Our approach is multi-stage. We first…

ModeSwitch-LLM is a lightweight request-boundary controller for improving single-GPU large language model inference efficiency by routing each request to an appropriate fixed inference mode. Instead of relying on one static serving…

Low energy barrier magnet (LBM) technology has recently been proposed as a candidate for accelerating algorithms based on energy minimization and probabilistic graphs because their physical characteristics have a one-to-one mapping onto the…

We demonstrate a new technique that adapts single-qubit randomized benchmarking to two-qubit M{\o}lmer-S{\o}rensen gates. We use the controllable gate phase to generate Cliffords that act on a two-state subspace, enabling benchmarking of…

Transformer-based large language model (LLM) inference serving is now the backbone of many cloud services. LLM inference consists of a prefill phase and a decode phase. However, existing LLM deployment practices often overlook the distinct…

Large Language Models (LLMs), with billions of parameters, present significant challenges for full finetuning due to the high computational demands, memory requirements, and impracticality of many real-world applications. When faced with…

Safety benchmarks are routinely treated as evidence about how a language model will behave once deployed, but this inference is fragile if behavior depends on whether a prompt looks like an evaluation. We define evaluation-context…

Recently, Anil et al. (2024) show that many-shot (up to hundreds of) demonstrations can jailbreak state-of-the-art LLMs by exploiting their long-context capability. Nevertheless, is it possible to use few-shot demonstrations to efficiently…

Detecting tricky bugs in plausible programs, those that pass existing test suites yet still contain bugs, remains a significant challenge in software testing. To address this problem, we propose TrickCatcher, an LLM-powered approach to…

Single-turn safety evaluation is a poor proxy for real fraud defense, where attackers escalate across multiple rounds. This paper evaluates fraud defenders under replay and adaptive multi-round attacks and measures when a defender refuses,…

Recent innovations in generative large language models (LLMs) have made their applications and use-cases ubiquitous. This has led to large-scale deployments of these models, using complex, expensive, and power-hungry AI accelerators, most…

Large language models (LLMs) are increasingly needed for interactive mobile applications, but high-quality models exceed the limited DRAM available on smartphones. Flash storage can hold larger models, yet flash-backed inference is slow…