Knowledge catalogue
arXiv:2606.23671v2 Announce Type: replace Abstract: Prior work shows that large language models (LLMs) exhibit introspective capability on benign tasks. We extend the question to safety contexts and e
arXiv:2606.29689v1 Announce Type: new Abstract: Open-ended aesthetic critique is a challenge for multimodal large language models (MLLMs): unlike multiple-choice aesthetic benchmarks, it has no single
arXiv:2606.30454v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as agents in simulations of social systems, yet it remains unclear when their behavior can be inter
arXiv:2606.29088v1 Announce Type: cross Abstract: There are various benchmarks to evaluate bugfixing capabilities of Large Language Models. However, most widespread benchmarks do not fully reflect rea
arXiv:2606.30256v1 Announce Type: new Abstract: Safety benchmarks often buy scalability by fixing the prompt, the language, and the turn structure. For emotional-support chatbots, that bargain hides p
arXiv:2602.19778v4 Announce Type: replace-cross Abstract: Automatic Chord Recognition (ACR) is constrained by the scarcity of aligned chord labels, as well-aligned annotations are costly to acquire. A
arXiv:2606.30220v1 Announce Type: new Abstract: High benchmark accuracy does not guarantee genuine use of visual evidence. We study this problem in traffic accident Video Question Answering (VideoQA),
arXiv:2606.29315v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to take actions in the real world and support human decision-making, yet most agents rely on parametr
arXiv:2602.15257v3 Announce Type: replace-cross Abstract: We present the first comprehensive, large-scale study of training long-context vision language models up to 344K context, targeting long-docum
arXiv:2606.30393v1 Announce Type: new Abstract: Photographs frequently contain visual distractors besides foregrounds and backgrounds of the intended subject, competing for attention and weakening com
arXiv:2606.29520v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used as assistants across the software development lifecycle, yet their ability to reason about software
arXiv:2606.30573v1 Announce Type: new Abstract: We introduce SWE-Interact, a new testbed for evaluating coding agents on multi-turn, interactive, user-driven software engineering tasks. Existing front
arXiv:2606.29278v1 Announce Type: new Abstract: We introduce the Complexity Ceiling Benchmark (CCB), a controlled evaluation of how language-model reasoning decays as the number of required sequential
arXiv:2606.13079v2 Announce Type: replace-cross Abstract: Nowadays, the autonomous execution of cyberattacks capable of causing substantial real-world harm is widely regarded as one of the critical re
𝗚𝗟𝗠-𝟱.𝟮 (the latest open weights model) is having an Enterprise moment, and it is not an exaggeration.🚀 🔥 We have been impressed by how strongly GLM-5.2 is pushing long-horizon performance .. not just
arXiv:2606.29034v1 Announce Type: cross Abstract: Large language models (LLMs) increasingly summarize clinical evidence, where a claim's weight depends on how strongly it is supported. Yet these model
arXiv:2606.29054v1 Announce Type: new Abstract: Large language models (LLMs) deployed for structured generation (NER, JSON extraction, QA, and classification) lack formal reliability guarantees, and s
arXiv:2606.30383v1 Announce Type: new Abstract: A rapidly growing class of LLM agents is multi-party: the agent acts for a principal (who briefs it, sends follow-ups, and receives results) while also
arXiv:2504.16116v4 Announce Type: replace-cross Abstract: The Web3 ecosystem, underpinned by cryptographic primitives and decentralized consensus, represents a high-stakes environment where software v
arXiv:2606.27731v1 Announce Type: cross Abstract: Despite their strong general capabilities, large language models (LLMs) often remain unreliable when outputs must be numerically precise. A key reason
arXiv:2606.27378v1 Announce Type: new Abstract: We introduce an axiomatic evaluation framework for latent thought representations in LLMs, comprising metrics that are independent of downstream benchma
It's quite rational to regulate frontier API models, especially to get more transparency for the government, without regulating open-source AI. Here's why: 1. The most dangerous AI systems right now a
arXiv:2606.26479v1 Announce Type: cross Abstract: Recent work (2024 to 2026) has converged on a strategy for defending tool-using LLM agents against indirect prompt injection: rather than training the
arXiv:2606.27251v1 Announce Type: cross Abstract: Building persistent embodied agents in unstructured environments demands unified orchestration of heterogeneous tools spanning both cyber (APIs, IoT)
arXiv:2606.26541v1 Announce Type: new Abstract: Data from affected populations are crucial for informing humanitarian response, but their value depends on timely and consistent interpretation of nuanc
arXiv:2606.26489v1 Announce Type: new Abstract: News media play a central role in shaping public perceptions of climate change, and whether coverage emphasizes threats or solutions has measurable effe
arXiv:2606.26936v1 Announce Type: cross Abstract: With a profusion of jailbreaks for LLMs now widely known, a growing concern is that non-expert malicious actors ('the average Jane') could elicit acti
arXiv:2606.27347v1 Announce Type: new Abstract: Whether political elites organise into rent-seeking coalitions that capture public resources or civic networks that sustain governance is a central ques
arXiv:2606.26493v1 Announce Type: new Abstract: Diffusion language models offer a promising alternative to autoregressive models due to their potential for parallel and iterative generation. However,
arXiv:2606.26654v1 Announce Type: new Abstract: Personalized language-model assistants are often evaluated through a memory lens: can a model recall preferences users have explicitly stated in dialogu
arXiv:2606.26987v1 Announce Type: cross Abstract: Recent work identified emotion vectors in Claude Sonnet 4.5, which are internal representations that encode emotion concepts, causally influence behav
arXiv:2606.25375v1 Announce Type: new Abstract: With the rapid adoption of generative AI, synthetic medical images pose growing risks, including diagnostic deception and insurance fraud. Although prio
arXiv:2606.25432v1 Announce Type: cross Abstract: Inference efficiency is typically pursued by shrinking the model: distillation, pruning, quantization, and sparse routing each lower per-token cost wh
arXiv:2606.25013v1 Announce Type: cross Abstract: Today's reasoning models use thinking tokens to attain stronger performance on benchmarks than their instruction-tuned counterparts. It is also genera
arXiv:2606.25491v1 Announce Type: new Abstract: Automated homework assessment depends not only on recognizing student answers, but also on accurately locating where each answer and each intermediate r
arXiv:2606.25092v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) models route each token to a few of many experts, inviting the hypothesis that experts form functional modules tied to c
arXiv:2606.26079v1 Announce Type: new Abstract: Standard benchmarks for multimodal large language models (MLLMs) score each item on one canonical ordering and miss whether order-irrelevant shuffling c
arXiv:2606.24083v1 Announce Type: cross Abstract: 'Talk short. Drop grammar. Save token.' This caveman style is widely promoted as a way to cut inference cost, but whether it actually saves anything d
arXiv:2606.24422v1 Announce Type: new Abstract: We introduce EgoSAT, the first comprehensive benchmark for egocentric video reasoning in streaming settings, designed to evaluate the capabilities of mo
arXiv:2606.24829v1 Announce Type: new Abstract: Camera-prompted text-to-video (T2V) models are increasingly used to synthesize virtual camera captures, such as orbiting objects or moving through stati
arXiv:2606.24335v1 Announce Type: new Abstract: Counterfactual analysis is widely used to study evidence use in vision-language models, but its diagnostic value is limited on well-posed tasks: when se
arXiv:2606.24381v1 Announce Type: cross Abstract: Prompt-based interaction has become a dominant paradigm for using large language models (LLMs), where multiple candidate prompts are evaluated and the
Unlimited OCR is a great model on table parsing and understanding proper reading order. However it does struggle a little on semantic formatting, charts (it does decent at bounding boxes). Attaching t
We've provided some updated results on Mistral OCR that make use of the annotation feature for charts. The overall score is ahead of GPT-5.5 and just behind Gemini 3.1 Pro, which is quite impressive f
arXiv:2606.21645v1 Announce Type: cross Abstract: Large language models (LLMs) can readily reproduce conventional expressions, yet their ability to model gradient frequency distributions remains under
arXiv:2606.22000v1 Announce Type: cross Abstract: We introduce CFAgentBench, a reproducible, self-hostable environment and benchmark for autonomous construction-finance agents: a CFO/controller-class
arXiv:2606.21083v1 Announce Type: cross Abstract: Large language models (LLMs) deployed for logical reasoning in knowledge-intensive domains exhibit a subtle but critical failure: coherence can be vac
arXiv:2606.15980v2 Announce Type: replace Abstract: Activation monitors -- lightweight probes trained on a language model's internal representations -- are an increasingly common layer in deployment s
arXiv:2606.21249v1 Announce Type: new Abstract: Retrieval heads, attention heads that copy information from earlier context to the current position, have been proposed as the mechanistic substrate for
arXiv:2606.23669v1 Announce Type: new Abstract: Text-to-image models can generate visually plausible city streets, but whether their outputs correspond to a requested road segment rather than a generi
arXiv:2606.19636v2 Announce Type: replace Abstract: Math and science reasoning benchmarks rely on pass@k, the fraction of sampled chains that reach gold, as the canonical per-example difficulty signal
GLM-5.2 is great at design (Opus level IMO). I am also starting to see great results with long-running tasks, too. How is this possible? I think there are a few clever hacks. But I just came across th
I have some very big news... KernelBench-Hard with H100 and B200 (single gpu results) AND KernelBench-Mega tested on RTX PRO 6000, H100, B200 is finally out! Starting with Mega, each of models wrote a
arXiv:2408.02600v3 Announce Type: replace Abstract: Background. Biomedical language models should improve performance on biomedical text while retaining general-language-modeling fluency. For Mamba-ba
arXiv:2606.11961v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as conditional generators for structured data, relying on in-context learning (ICL) to adapt to new
arXiv:2606.11289v1 Announce Type: new Abstract: Diffusion models have consistently driven progress in text-to-image generation. However, it is challenging to attribute recent progress to specific mode
As vertically integrated platforms start to dominate they lock out third party access to the most valuable portions of the platform. Of course, Anthropic is has the right to implement whatever policy
arXiv:2606.10142v1 Announce Type: new Abstract: Recent advances in 3D generation have led to substantial improvements in realism, controllability, and efficiency, yet the evaluation of 3D assets remai
arXiv:2603.09979v2 Announce Type: replace Abstract: Persian poetry plays an active role in Iranian cultural practice, where verses by canonical poets such as Hafez are frequently quoted, paraphrased,
arXiv:2606.09850v1 Announce Type: cross Abstract: Post-training alignment algorithms are predominantly evaluated as black boxes, obscuring how they reshape language models' internal computations. We p