Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2609.14124v1 Announce Type: cross Abstract: Medical image analysis is often hindered by biased datasets, which can lead to biased models and limited clinical applicability. A promising strategy...
Chemical reasoning language models are expected to produce faithful chain-of-thought (CoT) explanations when answering chemistry tasks, but across four model families and twelve tasks, hallucinations are widespread and largely independent of answer correctness. Attribution analyses reveal that these models use a shared scratchpad function: Chem‑R and ether‑0 rely on fragmented SMILES drafts, while ChemDFM‑R emphasizes scaffold, positional, and naming cues. Perturbing Chem‑R’s SMILES sketches degrades generation, indicating that structural drafts can be causally load‑bearing even when verbal structural claims are largely inert.
arXiv:2609.14759v1 Announce Type: cross Abstract: Alignment applied after pretraining is shallow in a measurable way: a single direction in a model's residual stream can be edited out, and the model...
arXiv:2609.14761v1 Announce Type: cross Abstract: Visual feedback in the form of user-uploaded images and videos is becoming increasingly common in e-commerce platforms because it provides authentic...
arXiv:2609.14795v1 Announce Type: cross Abstract: There is a tradeoff in machine translation meta-evaluation between prioritizing alignment with adequacy versus fluency. The balance depends on the co...
arXiv:2609.14839v1 Announce Type: cross Abstract: The integration of Large Language Models (LLMs) into automated code optimization introduces a critical reliability risk we term the Efficiency Halluc...
arXiv:2609.15606v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have achieved remarkable progress on short video understanding yet remain limited on long videos due to the...
arXiv:2609.15625v1 Announce Type: cross Abstract: Driven by the escalating demand for privacy-preserving computing, Federated Learning (FL) has witnessed remarkable progress, becoming a cornerstone t...
arXiv:2609.15671v1 Announce Type: cross Abstract: Visual Question Answering (VQA) with Vision-Language Models (VLMs) is increasingly used in privacy-sensitive and bandwidth-constrained settings. Fede...
arXiv:2603.13359v2 Announce Type: replace Abstract: Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusa...
arXiv:2503.06868v2 Announce Type: replace-cross Abstract: Existing long-text generation methods produce lengthy outputs from short inputs, leaving long-input-to-long-output generation underexplored....
arXiv:2503.17894v4 Announce Type: replace-cross Abstract: We propose a generative learner for estimating conditional average treatment effects and characterizing the full distribution of these effect...
The paper introduces CORAL, a multimodal framework that combines spatial grounding and concept-level supervision for medical report generation. CORAL uses a prompt-driven segmentation model to localize lesions and a Concept Bottleneck module to predict multi-class clinical attributes, feeding these textual concept tokens and mask-modulated visual features into a multimodal large language model. Experiments on BUS-CoT and IU X-ray datasets show that CORAL improves diagnostic accuracy, concept consistency, and report quality compared to existing general-purpose and medical MLLMs.
arXiv:2508.02312v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs), now a foundation in advancing natural language processing, power applications such as text generation, machine...
arXiv:2602.08370v2 Announce Type: replace-cross Abstract: Realizing versatile and human-like performance in high-demand sports like badminton remains a formidable challenge for humanoid robotics. Unl...
arXiv:2602.13576v2 Announce Type: replace-cross Abstract: Evaluation and alignment pipelines for large language models increasingly rely on LLM-based judges, whose behavior is guided by natural-langu...
arXiv:2603.19308v2 Announce Type: replace-cross Abstract: In autonomous driving, multi-agent collaborative perception enhances sensing capabilities by enabling agents to share perceptual data. A key...
arXiv:2605.22902v2 Announce Type: replace-cross Abstract: Generative Vision-Language Models (VLMs) perform well on multimodal reasoning, but how visual inputs are transformed to text remains poorly u...
arXiv:2609.13199v1 Announce Type: new Abstract: Knowledge distillation aims to improve the performance of lightweight student models by transferring knowledge from larger and more powerful teacher mo...
arXiv:2609.14778v1 Announce Type: new Abstract: We propose a decentralized privacy-preserving learning algorithm in which each agent holds a single private sample and a shared model. Samples are lear...