All papers

Evaluates the safety of ChatGPT health advice using real user conversation logs across four Global South countries.

The paper argues that software agency should be treated as a core organizing abstraction in computer science rather than merely an AI application.

Analyzes answerability signals in frozen visual question answering models, finding standard confidence and alignment metrics fail to reliably detect unanswerable inputs.

This paper compares the quality and coverage of unit test cases generated by ChatGPT against an automated generation tool.

The paper presents an oversampling method using a large language model to improve malicious JavaScript code detection.

The paper introduces an image augmentation pipeline that generates diverse visual variants to anchor multimodal chain-of-thought reasoning in visual evidence.

The paper introduces a benchmark assessing whether merchant websites provide adequate access, terms, and evidence for autonomous shopping agents.

This paper presents an optimization method that trains models to decide when to invoke reasoning tokens, reducing unnecessary thinking overhead.

Presents Interaction-as-a-Judge, using computer-use agents to evaluate interactive web and desktop software applications through direct dynamic interaction.

Proposes a reinforcement learning framework that tracks external versus internal knowledge sources to reduce search agent fabrication.

The paper presents an encoder-only transformer ranking model deployed at Disney+ that uses lightweight debiasing to improve recommendations.

The paper proposes an agent harness that autonomously evolves questioner and solver workflows from unlabeled video data without updating base model parameters.

Analyzes how multi-agent workflow architecture affects the propagation and attenuation of coordinated stance manipulation across sequential agent chains.

Introduces ProtoArena, a black-box dynamic benchmark evaluating the ability of AI agents to discover and exploit cross-component network protocol vulnerabilities.

This paper explores how consensus planning methods affect the code generation capabilities of language models.

Introduces an instruction-following benchmark and training recipe to improve goal adherence in full-duplex conversational models.

The paper presents a multi-agent framework that improves code generation by planning programming blueprints and repairing generated code.

The paper evaluates carbon-aware scheduling for geographically distributed language model serving under strict latency service-level objectives using an empirical simulator.

The paper presents a predictive KV cache quantization method exploiting temporal predictability to reduce memory overhead in autoregressive video generation models.

Introduces LatentRiskBench to evaluate safety risks in enterprise agents where action safety depends on latent workflow context.

SWEPatchTest provides a benchmark to evaluate whether agent-generated test suites correctly accept valid patches and reject invalid ones.

Evaluates public LLM watermark detectors and introduces a split-key mechanism combining public detection with private verification.

This paper introduces a framework that acquires and directly executes reusable visual programs to improve multimodal reasoning agent reliability.

The paper introduces an open-source framework and recipe for training and evaluating deep-search language agents on complex web exploration tasks.

The study evaluates scientific hypotheses generated by language models, concluding that they extend existing knowledge rather than starting new research directions.

The paper introduces action chunking for agentic robotics to execute pre-planned primitive routines without redundant model reasoning at each step.

Proposes AutoReview, a multi-agent system combining retrieval-augmented generation and code slicing for security-focused code review.

Proposes a security meta-model mapping attack surfaces, vulnerabilities, and risks across retrieval-augmented generation pipelines.

The paper proposes a distillation framework to remove specific multimodal knowledge without causing hallucinations or leaving residual queryable information.

The paper examines how providing additional context, such as codebases and benchmark transcripts, affects automated LLM safety auditors.

The paper introduces a benchmark assessing whether merchant websites provide adequate access, terms, and evidence for autonomous shopping agents.

The authors evaluate how 8-bit and 4-bit quantization methods affect model accuracy across Bavarian, Standard German, and English benchmarks.

This review surveys research on using language models for code verification and assessing the safety of model-generated code.

The paper analyzes vulnerabilities from persistent artifacts shared across AI agent skill sessions and develops a runtime defense against resulting cross-session attacks.

Introduces a benchmark probing physical perception tasks related to humans, including depth, trajectory, illumination, and kinematics.

The paper examines how shared KV caches cause language model agents to leak goals across competing clients, offering internal intervention techniques.

Introduces LatentRiskBench to evaluate safety risks in enterprise agents where action safety depends on latent workflow context.

This paper introduces a self-distillation method that searches over repeated layer execution orders to create stronger teacher guidance.

The paper presents a tool that uses language models and static analysis to generate valid visual flowcharts from legacy codebases.

The paper conducts an empirical study tracking how publicly flagged vulnerable AI models propagate across code repositories and downstream applications.

This paper evaluates the ability of language models to generate non-functional requirements from functional software specifications.

Introduces VARI, a group-structured intent alignment framework to improve text-to-video generation robustness across equivalent prompts.

Shows agent failure detection models often rely on trivial cues like run length, exposing benchmark flaws when attributing critical failure steps.

The paper introduces an open-source framework and recipe for training and evaluating deep-search language agents on complex web exploration tasks.

Introduces PaceReasoner, using long reasoning outputs during test-time reinforcement learning to boost the accuracy of brief, low-latency model answers.

The paper meta-analyzes 25 years of automated verbal deception detection models and identifies a persistent accuracy ceiling around 70 to 75 percent.

This paper presents an empirical study evaluating open-source language models for automated unit test generation in software development.

This paper introduces a grounded alignment framework using anomaly detection datasets and patch-level policy optimization to fix distortions in flow generators.

This paper audits gender and stereotypical bias in multilingual job-matching embedding models using robust statistical tests.

Evaluates different agent control granularities for post-route FPGA timing closure to balance execution reliability and frequency gains.

The paper analyzes legal eDiscovery review data to compare generative AI tools against continuous active machine learning and manual human review.

Presents RT-J, a relational transformer pretrained on diverse databases to achieve strong in-context table predictions with few sample examples.

This paper presents the design and real-world developer evaluation of an automated code generation assistant for Ansible automation scripts.

TableVision recovers verified table structure from historical documents using dynamic programming and column sum checks.

Analyzes how persona construction in generative social simulations causes network segregation and collective political polarization.

The paper examines how providing additional context, such as codebases and benchmark transcripts, affects automated LLM safety auditors.

The paper investigates how fine-tuning language models on descriptions of user motivations influences their downstream trust and cooperative behavior.

LatentRemedy uses task-conditioned latent tokens to allow safe key-value cache reuse across conflicting tasks in multi-agent systems.

This paper presents an agentic framework to compile large multi-modal requirement documents into runnable software systems using test-driven workflows.

This paper presents a verification and recovery framework using vision-language models to correct tracking errors and mask drift in video object segmentation.