All papers

The paper evaluates whether language models genuinely benefit from compressed memory by comparing them against memory-free baseline models.

This paper benchmarks language model agents against skilled humans on binary software reverse engineering tasks using interactive tools.

This paper introduces a benchmark to measure whether autonomous coding agents truly improve their policies from interaction traces.

The authors introduce Bemula, a benchmark designed to evaluate how accurately language model agents erase specific user memories upon request.

The paper uses activation steering of affective circuits in vision-language models to increase hazard-detection recall for post-disaster damage assessment.

Applies convolutional neural networks and vision transformers to video feeds for automated detection of labor contractions in dairy cattle.

Proposes an interpretability framework using sparse autoencoders to explain mixture of experts routing based on superposed fine-grained features.

Analyzes security incidents from frontier AI agent evaluations and proposes a framework for proactive boundary assurance.

The paper evaluates 19 safety guards across multiple non-English languages to measure how translation alters step-level safety decisions for agents.

The paper proposes a multi-view hierarchical memory framework to improve information retrieval from enterprise group chats.

The paper studies compute allocation across multi-agent coding teams to identify Pareto-optimal trade-offs between cost and performance.

The paper studies how much models contribute to multi-model teams, showing individual test scores do not correlate well with collaborative performance.

The paper evaluates whether contrastive encoder adversarial alignment metrics transfer to target recovery by independent models across different perturbation budgets.

Proposes a lightweight fruit sorting system combining 3D convolutions and Mamba sequence modeling for rotational visual quality inspection.

Presents a hypernetwork that maps document contexts directly into low-rank adapter weights for large language models.

Audits the reliability of LLM judges across multiple prompt formats, temperatures, and orderings, introducing a trustworthy verdict metric.

The paper analyzes why search-augmented language models produce outdated answers even when relevant evidence is retrieved.

The paper shows mean benchmark accuracy misses tail risks in fine-tuned models, proposing recipe certification cards for deployment decisions.

The paper proposes an agent framework that discovers lightweight proxy evaluators to bypass costly benchmark evaluations during autonomous engineering discovery.

The paper presents a taxonomy that locates agent workflow failures across specific workflow stages and agent functional capabilities.

Fine-tunes SAM 2.1 on multi-class building images to improve 2D segmentation for image-based 3D reconstruction.

Introduces a benchmark evaluating how misleading visualizations and charts affect the multi-step execution and decisions of web agents.

Implements open-vocabulary speech recognition on a low-cost ESP32-S3 microcontroller without dedicated neural accelerators.

Evaluates how schema ordering impacts LLM performance under structured output constraints, showing reasoning-first designs prevent accuracy drops.

The paper studies how semantic layers, orchestration, and compaction impact the performance and behavior of data agents across benchmarks.

The paper audits multimodal clinical machine learning studies, finding that incomplete unimodal baselines frequently inflate claimed gains.

The paper evaluates judge labels on agent steps by restarting failed trajectories from intermediate states to identify true failure points.

The paper introduces an imperceptible image perturbation attack that manipulates alignment in vision-language models during preference optimization without altering text annotations.

Introduces an architecture with teacher-specific tokens and mixture-of-experts to scale multi-teacher distillation in digital pathology.

Develops a GPU-accelerated tokenizer using CUDA graphs to reduce time-to-first-token in agentic language model serving.

Evaluates whether LLM compliance systems adjust verdicts when governing rules are perturbed, finding many verdicts are insensitive to rule changes.

Introduces a benchmark to evaluate how effectively LLMs clarify ambiguous user requirements in repository-level software development tasks.

Introduces a compact language model trained with distillation and reinforcement learning to manage long-term memory across multiple external frameworks.

The paper introduces a benchmark to study why frontier models prioritize protecting subjects over preventing larger physical property damage.

Introduces Bifocal, a training-free mixed-granularity visual token pruning method for multimodal models to protect fine-grained task accuracy.

The paper analyzes the failure modes of agent memory reuse and evaluates certified memory update strategies when facts change.

The paper presents CELL, learning pose-informative correspondence confidence end-to-end via differentiable probabilistic PnP to localize event cameras in LiDAR maps.

Proposes a method to select a compact subset of procedural skills that provide effective training signals during model distillation.

Presents an execution-based evaluation pipeline using browser agents to verify language model solutions for no-code software bug reports.

Evaluates typed decision models used as agent guardrails, uncovering vulnerabilities and asymmetric error rates in allow-or-block actions.

Presents a benchmark that measures long-context reasoning difficulty using evidence scope, dispersion, and reasoning depth.

The authors evaluate how language models alter their refusal behavior when informed they are being monitored in English and five Indic languages.

The paper tracks automatic speech recognition performance from 2021 to 2025, finding persistent accuracy disparities between standard American English and minority accents.

The paper presents CWE-bench, a benchmark of 120 held-out software tasks to evaluate AI agents on finding and repairing security vulnerabilities.

Introduces a post-training framework that trains text-to-SQL agents directly inside interactive database execution environments.

Develops an open-source testbed to evaluate segment-wise key-value cache reuse beyond exact prompt prefixes during language model serving.

Introduces a benchmark evaluating guard models on identifying unfulfilled safety obligations in language model agent trajectories alongside forbidden actions.

Deploys a quantized YOLOX-S detector on embedded hardware for onboard maritime satellite imagery selection and data reduction.

The paper analyzes how answering via a separate model call improves evidence adherence when reasoning traces conflict with provided context.

The paper analyzes integrity risks in deployment workflows when secondary repository files change without modifying the approved primary configuration file.

The paper derives multimodal data mixing scaling laws across text, vision, and audio using mixture-of-experts models.

Shows that language agents often ignore identified cross-domain dependencies, causing unintended disruptive actions in network management.

The authors introduce a benchmark to evaluate whether personal agents can distinguish accepted user arrangements from unacknowledged proposals.

The paper presents a diffusion planner with causal factorization for precision multi-object non-prehensile placement using mobile robots.

Measures how easily user-created web posts become cited sources in commercial AI search engines.

The paper shows that whole-context KV cache rehearsal harms compression under tight budgets, proposing targeted rehearsal to preserve answer-relevant entries.

Presents a benchmark with human-annotated visual evidence to evaluate retrieval and question answering across long documents.

The paper presents an in-stream monitoring tool that detects security vulnerabilities and prompts repairs during reasoning-model code generation.

The paper presents a repair method that re-encodes causal suffixes to enable safe non-prefix KV-cache reuse in retrieval-augmented generation.

The paper introduces a benchmark to evaluate whether language-model guardrails properly distinguish authorized from unauthorized actions across identical operations.