LLM 理论进展

Autoregressive large language model (LLM) decoding re-reads a growing key-value (KV) cache at every step, so long-context attention is bound by graphics processing unit (GPU) memory bandwidth. Block-sparse attention skips low-contribution KV blocks, but a selector that decides after the current query-key (QK) product, such as max-relative block thresholding (BLASST), still reads every K block, and a processing-in-memory (PIM) filter that decides from the current query places a serial PIM stage on the critical path. We present AMEND, a GPU-PIM attention design that removes both dependencies. AMEND predicts each block's BLASST verdict from margins audited at earlier steps, so the GPU fetches only predicted survivors while near-bank PIM units in high-bandwidth memory (HBM) concurrently score the omitted complement. A stack-level controller merges both observations, updates the predictor, and eagerly generates the next step's mask, so every predicted drop is re-observed without blocking the current token. Operating points are selected offline by constrained Bayesian optimization under a false-drop budget. Across LongBench and RULER runs, AMEND preserves near-baseline task quality; in simulation at batch size 8, it achieves $1.40$-$3.63\times$ end-to-end decode speedup and 28-66% lower dynamic decode energy than dense attention across 8K-64K contexts.

Depth pruning reduces large language model (LLM) inference cost by removing complete Transformer blocks. Activation-based methods collect hidden states through forward passes on calibration data, while existing forward-free methods score each Transformer block separately without measuring similarity between blocks. We propose Weight-Redundancy Pruning (WRP), a forward-free depth-pruning method that estimates inter-layer redundancy from checkpoint weights to select blocks without calibration data or model forward passes. WRP compares attention output and MLP down-projection weights across layers and combines their pairwise similarities with relative projection-scale information. The resulting all-pairs similarity matrix guides layer grouping and block selection. Across multiple pruning settings, model families, and downstream tasks, WRP consistently outperforms existing forward-free magnitude pruning and approaches the performance of activation-based methods.

Large language models often repeatedly select the same answer even when many alternatives are plausible. Prior work treats this concentration as bias, but it does not distinguish stable model preferences from responses that depend on a particular prompt wording. We introduce a bias depth score that measures both how strongly a model prefers its top answer under direct prompting and whether that answer survives scenario reframing. Across 4,442 opinion prompts and four large language models, only about a quarter of the concentrated preferences survive reframing. We call these persistent cases Deep biases, and the remaining prompt-dependent cases Shallow biases. Our results show that Deep biases are more often inherited from pretraining and preserved through SFT. Under both continued fine-tuning and prompt-based debiasing for diversity, Deep biases are consistently harder to remove than Shallow biases. Bias depth therefore separates stable learned biases from prompt-wording artifacts that single-prompt metrics conflate. Code, models, and data are available at deepbias.github.io.

Personalized language models aim to adapt responses to individual users, whose preferences are often latent and revealed gradually through interaction. Existing training-free methods rely on stored histories or retrieved memories, but they often struggle to reconcile long- term preferences with short-term topic-specific needs. To address this issue, we propose HyperTrace, a training-free framework that formulates online personalization as latent preference tracing. HyperTrace maintains interpretable natural-language hypotheses over short-term intent and long-term preferences, and updates them through an SMC-style reweight process using an LLM-based surrogate choice model. By updating these hypotheses across turns and sessions, HyperTrace enables personalization without parameter updates. Experiments on PRISM and PersonaMem-v2 show that HyperTrace improves response alignment, preference prediction, and profile consistency over strong online baselines, demonstrating the effectiveness of tracing latent user preferences for robust personalization. Code and scripts are available in the repository: https://github.com/jiseshen/HyperTrace.

Large language models (LLMs) are increasingly deployed in globally used assistants, yet their default choices in culturally grounded everyday situations can systematically favour some cultures over others, affecting localisation, user trust, and equitable behaviour. Existing cultural benchmarks evaluate accuracy against a single "correct" answer, making it difficult to characterise an LLM's cultural preference prior when multiple culturally grounded responses are all valid; they also conflate default preferences with context-driven adaptation. We propose DiSCo, a distribution-first forced-choice evaluation framework that isolates default cultural priors and tests steerability via a four-level context gradient (C0--C3). Using DiSCo-Bench (304 items) derived from BLEnD spanning 12 cultures, we evaluate six diverse instruction-tuned LLMs. Default priors are heavily concentrated, with UK and US together absorbing approximately 35% of all selections despite representing only 2 of 12 cultures. Most critically, prompt-based steering consistently widens the selection gap between high- and low-resource cultures, and injecting explicit cultural facts produces negligible distributional disruption, confirming that cultural preference bias cannot be resolved through prompt-based personalisation alone.

Open-source large language models (LLMs) are increasingly competitive with closed-source models while offering transparency and the ability to run inference without exposing user inputs to a service provider. However, running large-scale models locally requires substantial computational resources. In practice, users may still resort to a third-party provider, giving rise to privacy and correctness concerns. Existing solutions that address these problems often impose substantial server overhead or introduce additional trust assumptions. In this paper, we present Maverick, a novel approach to private and verifiable LLM inference based on a protocol for delegating matrix-vector multiplication, a dominant operation in LLMs. At its core, Maverick provides, to our knowledge, the first information-theoretically sound verification protocol for matrix-vector multiplication delegation with transparent preprocessing, efficient (batch) verification, and virtually no server overhead. We combine this verification primitive with LPN-based pseudorandom masking to provide input privacy. We implement our matrix-vector delegation primitive and use it to build an end-to-end prototype of Maverick, which we evaluate on Qwen3-4B by measuring throughput in tokens per second. We evaluate client configurations with 1-8 threads. With one client thread and a CPU server using up to 128 threads, Maverick achieves throughput gains over local inference of up to 17x when privacy masks are generated online, 45x when they are precomputed, and 44x when only verification is required. With four client threads, the corresponding gains are 13x, 18x, and 17x. When server computation is no longer the bottleneck, client-side microbenchmarks with simulated network delay show speedups of 12x-20x, 34x-135x, and 38x-157x.

Fine-tuning large language models (LLMs) on domain-specific data is essential for downstream adaptation. In many deployments, a participant cannot hold the complete model locally. This happens because the model owner keeps the full model proprietary, or because the participant lacks sufficient compute resources. Split Learning (SL) addresses this by partitioning the model between the participant and a server so that only a small portion runs locally. When the underlying data is additionally distributed across multiple institutions with privacy requirements, Federated Learning (FL) further enables collaborative training across participants by sharing only model updates instead of raw data. In this combined setting, each client transmits intermediate activations to the server, and for LLM fine-tuning, this exchange poses an inherent privacy paradox. The autoregressive nature of LLMs causes the transmitted activations to leak the input, and existing perturbation-based defenses are fundamentally ineffective in this setting. We address this leakage through a learned obfuscate-and-recover scheme that protects participants' private datasets while still allowing an independently deployable model to be trained on the server side. Experiments demonstrate that our approach achieves strong privacy protection with modest utility loss and system overhead, making split-based federated LLM fine-tuning practically viable.

Large language models (LLMs) have become ubiquitous in software development, with LLM-based automated program repair tools increasingly used during code review. In this report, we explore the iterative blind use of LLMs as bug-fixers. Across multiple models and repair environments, we find that LLMs consistently claim to detect bugs in entirely bug-free programs while the rate of repair of buggy programs is less than that of the damage to correct programs. We also explore the long-term dynamics of this iterative process, and find that this frequently reaches a pseudo-bug-fixing cycle where the same changes are added and removed again ad infinitum. Lastly, via mechanistic probing, we unveil the existence of a steering vector which controls the editing propensity, suggesting that LLMs have an internal representation of ``buggy code", and that this representation is what is falsely activated to induce pseudo-bug fixing. These results provide insight towards the dynamics of fully autonomous bug-fixing systems, as well as stopping conditions under ambiguous goals.

The advance of multimodal large language models (MLLMs) has fundamentally reshaped the paradigm of human-computer interaction, especially speech interaction models capable of seamless conversations. Despite remarkable performance as general voice assistants, their performance in specialized domains remains underexplored, particularly in scientific areas. Scientific interactions introduce formidable challenges, involving rare technical terminology, spoken norms of abbreviations, and the natural verbalization of symbolic special expressions. In this paper, we introduce S$^3$-Bench, a systematic evaluation framework covering 10 major disciplines, consisting of a Knowledge set for speech question-answering and a Dialogue set for multi-turn progressive interactions with simulated user agents. By decomposing a complete atomic turn into stages of speech recognition, perception, knowledge utilization with reasoning, and response pronunciation, we systematically characterize the common challenges and performance tradeoffs of existing approaches. Furthermore, experiments on multi-turn interactions reveal persistent limitations in user adaptation and the generation of accurate, comprehensive, and efficient responses.

Deploying Large Language Models (LLMs) directly on mobile platforms at the edge is gaining traction due to a myriad of benefits, such as increased privacy, personalization, and reduced latency. However, LLMs have heavy computational requirements, which are difficult for resource-constrained mobile and edge platforms to fulfill. In addition to limited compute resources, mobile and edge systems often have a compact form factor and lack physical mechanisms to dissipate heat generated from high processor usage rates (e.g., fans) to prevent throttling and reduced processing power, which LLMs can easily cause. To mitigate these effects, prior works have proposed various power governing strategies, such as dynamic voltage and frequency scaling (DVFS), for reducing power and heat generation for heavy computational tasks on mobile platforms. Recently, DVFS methods tailored for mobile LLMs have also been proposed. However, these methods mostly focus on optimizing hardware parameters and processor frequencies, and they fall short under some thermally constrained scenarios. Drawing from recent advances in machine learning, we identify and take advantage of the key insight that not all tokens require full-depth inference to maintain high-quality generation. Motivated by this, we present PELM, a solution that augments traditional DVFS processor frequency tuning with two additional workload-specific knobs: 1) speculative decoding and 2) variable verification depth to expand the optimization space to multiple dimensions for more power efficient on-device LLM inference. In extensive evaluations across hardware platforms and datasets, PELM demonstrates superior performance compared to state-of-the-art power governing methods, with up to 23.1% speedup and 52.4% reduction in energy consumption, while maintaining comparable task performance. The source code is available at https://github.com/imec-nu/PELM.

Static analysis remains a cornerstone of software security, yet the effectiveness of tools such as CodeQL is often limited by the substantial manual effort required to develop high-coverage query suites. While large language models (LLMs) have emerged as a potential solution for automated code reasoning, their practical utility in generating structured, executable security queries remains underexplored. In this paper, we conduct an empirical study to evaluate the ability of LLMs to synthesize CodeQL queries using vulnerability data from the National Vulnerability Database. Through this investigation, we explore the potential of using LLMs as an automatic CodeQL query generator. Subsequently, we systematically evaluate the performance of various LLM architectures across a diverse set of real-world vulnerabilities, measuring their ability to improve detection coverage and precision. Our findings reveal that LLM-generated queries significantly enhance the baseline CodeQL queries, yielding 82% improvement in average F1-score. Furthermore, we provide a detailed cost- benefit analysis showing that while direct LLM-based scanning of entire repositories is often computationally and financially prohibitive, leveraging LLMs to synthesize CodeQL queries offers a scalable and cost-effective alternative for large-scale vulnerability detection. Our results suggest that LLMs can effectively bridge the gap between unstructured vulnerability reports and formal static analysis specifications, offering a scalable path toward comprehensive automated vulnerability detection.

Reasoning language models have made substantial advances on a variety of complex tasks, yet their capabilities remain overwhelmingly English-centric: models primarily reason in English regardless of the language they are prompted in. This is inaccessible for non-English-speaking users, risks losing the intent of the original question, and forgoes knowledge more readily expressed in the target language. In this work, we advance L2 reasoning, the ability of a model to reason consistently in the language of the user's prompt, thus building an in-language bridge between the prompt and the answer. We approach this problem from a data-centric angle, investigating how to optimize data composition and scheduling in SFT for reasoning generalization. Building Tiny Aya L2-Thinker at 3.35B scale, we achieve an L2 reasoning rate above 93% across 60 languages on 6 benchmarks spanning math, commonsense reasoning, instruction following, open-ended generation, and cultural reasoning while keeping performance strong. We show the path to generalizing L2 reasoning to held-out languages goes through broader language coverage, readily available multilingual non-reasoning data, and a sufficient English reasoning backbone. These findings indicate that reasoning is a language-agnostic behavior that can be transferred across typologically diverse languages through careful data mixing and without requiring reasoning supervision in every target language. We release our model weights and multilingual reasoning data to support further research on accessible, in-language reasoning.

Latent reasoning approaches enhance token-level efficiency and robustness by replacing verbose, explicit chain-of-thought (CoT) tokens with compact continuous-space embeddings. However, existing methods lack direct process supervision over these latent embeddings, which often leads to representation collapse and uneven information distribution. To address this, we propose Prototype-Mediated Process Supervision (PMPS), which introduces learnable reasoning prototypes as semantic anchors to provide structural process-level supervision for latent reasoning. PMPS projects latent embeddings and explicit CoT embeddings into a shared prototype space, achieving many-to-many soft alignment between unequal-length representations through prototype assignment. Meanwhile, we introduce a Progressive Sequential Alignment (PSA) module to further guide training: positional priors initially encourage sequential alignment structure, then gradually relax to permit adaptive matching. Experimental results show that PMPS compresses output token length to under 50% of explicit CoT on GSM8K-Aug. Compared to leading baseline SIM-CoT, our method achieves average accuracy gains of 2.08% across different model families. On GPT-2, PMPS even surpasses CoT-SFT. On larger models and a more challenging task, PMPS consistently attains the highest accuracy among all latent reasoning methods with comparable output length.

The development of generative artificial intelligence resources enables opportunities of speeding up systems and engineering design work. This contribution introduces a framework of formal operations for assembling context in LLM-based engineering design. This framework involves the assembly of modular context units, including policy prompts, reference units with persistence, and user questions with prompt vectoring. This approach enables the systematic structuring of interactions with generative models. A formal method for evaluating modelling-as-code LLM outputs is also presented, which enables the evaluation of compliance to intent from LLM answers and thereby asses the support from LLMs for systems architecture modelling.

Large language model safety and security research is preoccupied with, among other things, detecting and preventing jailbreak attacks: alignment bypasses that allow an adversarial user to elicit unwanted or harmful outputs from models. Arbitrary cipher, or covert communication, attacks are one such type of jailbreak and have previously been demonstrated against the fine-tuning APIs of commercial models. In these attacks, target models are trained on a corpus of encrypted harmful questions and responses and subsequently respond to harmful requests through the learned encryption scheme. In this paper, we show that newer frontier models do not require fine-tuning to acquire cipher-based communication skills. Instead, they can learn these skills through prompting and, when necessary, through in-context learning. Furthermore, model alignment is significantly weakened or entirely bypassed when communication occurs through the learned cipher. To the best of our knowledge, this constitutes a novel attack vector against commercial black-box large language models. We demonstrate successful jailbreaks against frontier models developed by Anthropic, Google, and OpenAI. Our attack bypasses commercial harmfulness classifiers because harmful content is encrypted and therefore appears as nonsensical text or gibberish.

In-context learning (ICL) is widely used in multimodal large language models (MLLMs) and achieves strong performance across a wide range of multimodal tasks. However, existing multimodal ICL methods often rely on surface level imitation of in-context demonstrations, making it difficult for MLLMs to align their responses with the reasoning path required by the given multimodal input. This limitation becomes more pronounced in complex multimodal tasks, thereby restricting further improvements in MLLM performance. To address this issue, we propose a new multimodal ICL framework that combines contrastive demonstration modeling with the self-refinement capability of MLLMs. Specifically, our framework reformulates each demonstration by explicitly contrasting a suboptimal response with a better response under the same input, together with a reasoning path that reveals how the response should be refined. This contrastive formulation makes the reasoning path toward the desired response more explicit and guides the MLLM beyond superficial imitation. Furthermore, because effective refinement depends on the current response, we introduce a response-conditioned retrieval mechanism to select demonstrations whose reasoning paths are more relevant to the current response. In addition, we use a lightweight alignment controller to predict response quality and determine whether further refinement is needed. Experiments on three types of multimodal tasks show that the proposed framework consistently improves MLLM performance, with particularly notable gains on visual question answering (VQA).

In-context learning (ICL) jailbreaks reveal a critical vulnerability in multimodal large language models (MLLMs): harmful demonstrations in the prompt can induce unsafe outputs without modifying model parameters. Despite extensive empirical evidence, existing work lacks a principled understanding of why such jailbreaks reliably succeed or how their effectiveness scales with context composition. We propose a posterior reweighting framework that models a safety-aligned MLLM as implicitly operating over competing behavioral modes, and interprets in-context demonstrations as inference-time evidence that dynamically shifts the model's posterior preference between safe and harmful behaviors. This view formalizes jailbreak as a process of evidence accumulation, yielding predictive scaling laws with respect to demonstration count, harmful ratio, adversarial strength, and semantic diversity. Guided by this framework, we introduce a posterior-aware inference-time defense that adaptively injects benign counter-evidence based on estimated risk, effectively suppressing harmful posterior drift while preserving model utility. Compared to existing in-context defenses, our method achieves a significantly improved robustness-utility trade-off under a fixed intervention budget. Together, our results establish posterior reweighting as a unifying and predictive framework for understanding and mitigating ICL jailbreak in MLLMs.

Large Language Models (LLMs) often generate natural-language comments while writing code, and these comments become part of the context used to generate the code that follows. However, it remains unclear which properties of comments affect code-generation performance. We study this question through observational analyses and controlled interventions. On LiveCodeBench, neither comment frequency nor broad comment intent reliably predicts pass@1. We then prefill weaker recipient models with comment blocks written by stronger source models, allowing us to separate comment surface form from the solution content they convey. Comments from source solutions that pass the tests raise recipient pass@1 by 17.2% on average. In contrast, comments describing failed solutions provide no reliable gain, while comments written for a different problem reduce pass@1 by 20.8%. Finally, across a wide range of models and prompt variants, most recipient models show no significant recovery of the external-comment gain, and the best case recovers only 24%. These results show that comments help code generation not merely because they are comments, but because they can provide correct solution content that prompting cannot reliably elicit.

Test adequacy criteria are widely used to evaluate and guide software testing. Although prior research has extensively examined these criteria using human-written programs, faults, and tests, the increasing adoption of Large Language Models (LLMs) for code generation raises important questions about their effectiveness in detecting LLM-induced faults. To investigate this, we conduct an empirical study involving 5 LLMs and 4 benchmarks, simulating end-to-end workflows in which both code and tests are automatically generated. We collect 6,000+ faulty program instances and evaluate the effectiveness and efficiency of 3 widely used adequacy criteria: statement coverage, branch coverage, and mutation testing. Our findings reveal several key insights. First, most faults introduced by LLMs are relatively trivial to catch. Second, the challenging faults are difficult to trigger using either traditional coverage-based or mutation-based criteria. Third, actual fault detection rates remain extremely low, often near zero, because test oracles fail to capture faulty behavior triggered by the generated test prefixes, exposing a critical limitation of automated test generation. Fourth, prompt-aware oracles can improve fault detection, but their overall effectiveness remains limited, highlighting the need for users to manually reason about test assertions. We further observe that mutation testing only marginally outperforms traditional coverage criteria in both triggering and detecting faults, raising questions about whether its significantly higher application cost is justified in this context.

Large language models have collapsed the cost of producing lexically elaborate prose, and whether peer reviewers still reward it is a question about the evaluator, not about the text. When the association between a writing cue and review scores moves across years, the reviewers may have changed, the submissions may have changed, or both, and a regression of scores on text cannot say which. We separate the two with a frozen rater: 81,850 machine reviews of ICLR submissions from 2018 to 2025, all generated in one February-April 2025 window with one model family and one prompt, so that its year-to-year coefficients track submission composition alone and the human-minus-frozen trend difference identifies reviewer preference drift. On 32,638 submissions with 124,615 human reviews, the human coefficient on non-domain lexical complexity falls from +0.142 to -0.015 while the frozen rater moves from +0.080 to +0.082; the three-way difference-in-differences is -0.0100 (q=0.013), and forty random-wordlist placebos through the same specification centre on zero. Humans still reward sentence-length variability, which the frozen rater never registers, while the frozen rater still pays for lexical complexity at its earlier rate. Every claim is held to a double gate of false-discovery control and interval exclusion, and the findings that failed adversarial re-testing are reported. Reviewers discounted a cue whose production cost collapsed, as models of manipulable signals prescribe; an LLM judge calibrated to historical human preferences inherits the earlier schedule and drifts out of alignment while its agreement with humans on totals stays ordinary.

Large Language Models (LLMs) are increasingly deployed with hierarchical instructions, yet they remain vulnerable to conflicts in which user directives override system-level constraints. Existing defense mechanisms predominantly focus on static input inspection and therefore fail to detect Response Drift, a phenomenon in which the model's final response violates system-level constraints despite seemingly compliant inputs. To bridge this gap, we introduce ELCD, a response-level latent conflict detector for post-generation, pre-delivery verification. Given the full generated output, ELCD constructs a composite hidden-state representation by concatenating the final-token embedding with the mean-pooled response embedding. It then optimizes a pairwise margin ranking objective to separate compliant and drifting responses in latent space. Extensive experiments across five mainstream LLMs ranging from 1.5B to 14B parameters demonstrate that ELCD significantly outperforms competitive baselines. Notably, it improves the PR-AUC on Llama-2-7B by approximately 30 percentage points and reduces the False Positive Rate at 95% TPR (FPR95) on Mistral-7B to 2.67%. These results suggest that ELCD provides a promising approach for latent instruction-conflict detection in open-weight or self-hosted LLM deployments.

Retrieved records are presentation units; a supplied partition determines which records enter a language model as one evidential contribution. We characterize the invariant group-content state that removes within-group copies while retaining complementary canonical content, show that equal group counts can encode different evidence states, and derive a sharp content-aware partition-error bound. Given a supplied partition, our pre-generation representation deduplicates and aggregates content within groups and bounds each group's contribution. Across 104,402 trials and 6 public checkpoints, a central natural-text intervention finds that content-fixed false splits add 10.27-32.66 percentage points and false merges remove 9.13-31.79 points; a matched six-slot control retains the positive direction in all 16 cells. In a new 48-item controlled campaign panel, changing the supplied partition produces measurable, checkpoint-dependent decision shifts across all four models, and the balanced mirror design exposes substantial order interactions. Together, the theory and experiments establish the supplied partition as a controllable pre-generation representation variable and characterize its checkpoint-dependent behavioral effects.

Retrieval-augmented personalization enables large language models to produce more accurate and preference-aligned outputs using relevant records retrieved from user histories. Personalized language models typically prepend a fixed number of retrieved user records, even when additional history is redundant, harmful, or unrelated to a user's distinctive behavior. We study minimal sufficient personalization: constructing the least costly ordered profile for each input while preserving the utility achievable from a retrieved candidate pool. We introduce ENOUGH, a method that iteratively appends behavioral records or emits STOP to construct profiles with adaptive lengths. Offline, bounded counterfactual search evaluates profile prefixes by jointly considering downstream gains, user specificity, and token costs. The resulting long-horizon targets are distilled into a multi-head value controller with explicit ranking and stopping supervision. At inference, the controller selects and orders records through lightweight decisions, and the frozen generator is invoked once after stopping. Extensive experiments on six personalized tasks demonstrate that ENOUGH consistently outperforms strong heuristic and retrieval-augmented baselines in both effectiveness and efficiency, achieving minimal sufficient profiles that preserve personalization utility while reducing unnecessary context costs.

The emergence of Chain-of-Thought (CoT) has established a robust foundation for Large Reasoning Models (LRMs). While deep reasoning is widely believed to enhance safety alignment, the stability of alignment mechanisms under extended reasoning remains underexplored. This paper challenges the prevailing view by revealing a critical vulnerability: Deep Reasoning May Induce Alignment Collapse. To rigorously quantify this phenomenon, we propose the Alignment Loss Rate (ALR) metric. Our experiments demonstrate that as reasoning depth increases, ALR rises significantly, indicating a severe degradation in model robustness against external perturbations. Capitalizing on this instability, a novel jailbreaking paradigm, Reasoning Trap (RT), is proposed. RT induces the model into extended reasoning to amplify the impact of adversarial attacks, leading to a sharp decline in safety capabilities. To elucidate the mechanism behind this collapse, we identify Attention Dilution as the root cause, arising from the competition for attention between the extended reasoning process and the original input. To mitigate this, Reasoning Residual Alignment (RRA), a lightweight defense strategy that dynamically re-emphasizes the input via residual connections integrated with the reasoning process.

Multi-step LLM reasoning lacks a machine-recheckable ledger: discarded reasoning paths leave no auditable record. We propose the Deposon scattering layer, which binds each node of an LLM-generated concept-decomposition graph to a two-parameter Deposon state; paths undergo three-channel scattering — transmission, reflection, irreversible dissipation — obeying T+R+A=1 for arbitrary parameters, with a maximum per-path energy-audit deviation of 2.2E-16 (machine epsilon). We report all three evidence tiers honestly. On synthetic trap benchmarks the path-filtering gain is closed (pre-registered): unified reaches 100% versus a decoy-capture baseline at 7%/10%. On real benchmarks the layer is indistinguishable from a trivial six-keyword rule filter (GSM8K 0.87 >= 0.85, McNemar p=0.5; StrategyQA 0.899 = 0.899); no difference is detected here, so we sharpen the claim to "the differential value lies solely in machine verifiability." Fusion yields a second negative result: convex combinations with a semantic prior never improve (physics 0.484 -> 0.452), and the apparent lambda=2 gain is an anti-field artifact; any fusion gain must be nonlinear. Modeling the reverse dynamics as a potential game on the graph, we evidence an auditable scalar's monotonicity and near-gradientness and quantify the empirical coordination ratio (ECR). The three formalized dynamical-equivalence propositions (P1a/P1b/T-P1c) are falsified under the pre-registered kill protocol, and the potential-game claim is downgraded to approximate (cyclic-graph median residual 0.669): only consistency-level evidence survives at the dynamical level. Code: github.com/zeroandcat/Deposon.

Software Fault Injection (SFI) supports testing of cloud systems by introducing software defects and observing their manifestation. Rule-based injectors such as ProFIPy provide controlled and reproducible source-level mutations but require fault patterns to be encoded manually. Large Language Models (LLMs) offer a data-driven alternative by generating context-dependent software faults. We compare two code LLMs, Qwen2.5-Coder and DeepSeek-Coder, with ProFIPy in OpenStack's Nova and Cinder services. On shared injection targets, activation and observable-failure rates are comparable, but operational profiles differ: LLM-generated faults produce more Catastrophic outcomes on Nova, whereas ProFIPy produces more Silent and Multi-component effects. The sampled LLM outputs also differ in how they manifest failure, while showing greater agreement in their propagation scope. These findings show that LLM-based fault injection extends the behavioral coverage of fixed fault models without establishing general superiority, and that practical adoption still requires controlled generation, runtime validation, system-level oracles, and reproducible experimental provenance.

Automatic fact-checking systems assess the veracity of claims given evidence from relevant documents. Large Language Models (LLMs) have demonstrated strong performance in fact-checking due to their general reasoning capabilities. However, it remains unclear whether they faithfully make use of the evidence provided to reach veracity judgments or rely on parametric knowledge. To investigate this, we introduce Fact-Ablated Evaluation (FAE), a new evaluation framework that iteratively ablates the cited evidence to assess whether LLMs revise their predictions accordingly. Our empirical results show that current off-the-shelf LLMs as fact-checking systems rely more on their parametric knowledge than on the evidence provided. To bridge this gap between prediction accuracy and evidence grounding, we propose REAL (Rigorous Evidence Ablation Learning), a training framework that promotes evidence-dependent verification through counterfactual evidence supervision for the LLM-as-verifier models. Experiments on four fact-checking datasets across different domains demonstrate that models trained with REAL obtain superior evidence-dependent capabilities compared to standard fine-tuned models. Our findings highlight that strong fact-checking performance can still coexist with weak evidence dependency, while REAL encourages veracity predictions to remain more closely tied to the availability of supporting evidence.

Firms making inventory decisions have access to operational data, optimization tools, and large language models (LLMs). Typically, data characterize the operating environment, optimization selects parameters within a prespecified inventory policy class, and LLMs support coding and decision analysis. We develop an integrated framework that combines these resources to automate inventory policy design. Given demand data, the framework iteratively uses an LLM to generate parameterized policy classes and an external solver to optimize its parameters within each class. Across 30 lost-sales inventory instances, the mean cost reduction relative to optimized base-stock benchmarks increases from 17.5% after one generation to 30.0% after ten generations. Parameter optimization is central to this performance: an LLM-only variant performs substantially worse, whereas optimization-guided feedback improves policy quality, accelerates search, and directs the LLM toward better policy classes rather than merely better parameter values within a fixed class. The strongest discovered policies are also interpretable: they combine recognizable inventory-control motifs, including capped orders, discounted or weighted pipeline inventory, and threshold-based replenishment logic. The search thereby produces new policy-class functional forms that, to our knowledge, have not previously been studied in the lost-sales inventory literature. These functional forms are not specified ex ante but emerge from the search process. Moreover, after their parameters are re-optimized, three discovered policy classes achieve average cost reductions of 21.75% to 22.60% across 10,064 new inventory instances. Overall, the results show that data-driven parameter optimization can guide LLM-based search over a broad space of inventory policy classes and identify high-performing, interpretable, and transferable decision rules.

Automated red-team attacks and blue-team defenses for large language models (LLMs) are advancing quickly. However, attackers and defenders are built and tested in isolation, and the resulting scores are hard to trust. To tackle this, we present ACEA (Adversarial Co-Evolution Arena), a platform that connects a pluggable red-team adapter and a pluggable blue-team adapter to a shared target LLM and scores their attack and defense rates with an LLM judge. ACEA contributes four components. First, a pluggable, model-agnostic arena. Any red or blue project connects over a minimal HTTP protocol, which we call the ACEA Standard Adapter Protocol (ASAP). It can be written in any language, and a project that exposes nothing but the protocol is a full participant. Second, an evaluation methodology built for adversarial rounds. Seeding the target with canonical secrets gives verifiable ground truth that separates real leakage from hallucination. We also send each attack to the target even when the defense blocks it, which measures the attack's raw potency independently of whether it was stopped. Together these yield a per-round decomposition of attack strength and defense effectiveness. Third, a real-time, game-style visualization with a detailed end-of-battle report that localizes each failure. The evaluation thus becomes an actionable signal for improving a red or blue project. Fourth, an optional in-context improvement loop that turns each round's outcome into advisory hints for the next. An adapter can then adapt across rounds without keeping state, provided it reads the hints. We describe the design of ACEA and the metrics through which red and blue teams are scored head to head.

Integrating robust safety guardrails into Large Language Models (LLMs) is essential for delivering helpful yet harmless responses. While proprietary systems exhibit reliable safety controls, their underlying methodologies and trade-offs remain largely undisclosed. Achieving comparable security in open-weight models remains a persistent challenge, as post-trained variants frequently suffer from over-refusal and degraded general quality. To overcome these drawbacks, we introduce Suan, a novel preference optimization algorithm. Unlike existing methods, we formulate the optimization objective directly at the gradient level, bypassing the standard variational derivation. As a result, we obtain more interpretable and robust training dynamics. Extensive evaluations across a diverse suite of competitive baselines and benchmarks demonstrate that Suan achieves superior safety alignment while fully preserving response utility.