As large language models (LLMs) are increasingly used in human-AI interactions, their social reasoning capabilities in interpersonal contexts are critical. We introduce SCRIPTS, a 1k-dialogue dataset in English and Korean, sourced from movie scripts. The task involves evaluating models' social reasoning capability to infer the interpersonal relationships (e.g., friends, sisters, lovers) between speakers in each dialogue. Each dialogue is annotated with probabilistic relational labels (Highly Likely, Less Likely, Unlikely) by native (or equivalent) Korean and English speakers from Korea and the U.S. Evaluating nine models on our task, current proprietary LLMs achieve around 75-80% on the English dataset, whereas their performance on Korean drops to 58-69%. More strikingly, models select Unlikely relationships in 10-25% of their responses. Furthermore, we find that thinking models and chain-of-thought prompting, effective for general reasoning, provide minimal benefits for social reasoning and occasionally amplify social biases. Our findings reveal significant limitations in current LLMs' social reasoning capabilities, highlighting the need for efforts to develop socially-aware language models.

We present Ring-1T, the first open-source, state-of-the-art thinking model with a trillion-scale parameter. It features 1 trillion total parameters and activates approximately 50 billion per token. Training such models at a trillion-parameter scale introduces unprecedented challenges, including train-inference misalignment, inefficiencies in rollout processing, and bottlenecks in the RL system. To address these, we pioneer three interconnected innovations: (1) IcePop stabilizes RL training via token-level discrepancy masking and clipping, resolving instability from training-inference mismatches; (2) C3PO++ improves resource utilization for long rollouts under a token budget by dynamically partitioning them, thereby obtaining high time efficiency; and (3) ASystem, a high-performance RL framework designed to overcome the systemic bottlenecks that impede trillion-parameter model training. Ring-1T delivers breakthrough results across critical benchmarks: 93.4 on AIME-2025, 86.72 on HMMT-2025, 2088 on CodeForces, and 55.94 on ARC-AGI-1. Notably, it attains a silver medal-level result on the IMO-2025, underscoring its exceptional reasoning capabilities. By releasing the complete 1T parameter MoE model to the community, we provide the research community with direct access to cutting-edge reasoning capabilities. This contribution marks a significant milestone in democratizing large-scale reasoning intelligence and establishes a new baseline for open-source model performance.

Large Language Models (LLMs) can reason over natural-language inputs, but their role in intrusion detection without fine-tuning remains uncertain. This study evaluates a prompt-only approach on UNSW-NB15 by converting each network flow to a compact textual record and augmenting it with lightweight, domain-inspired boolean flags (asymmetry, burst rate, TTL irregularities, timer anomalies, rare service/state, short bursts). To reduce output drift and support measurement, the model is constrained to produce structured, grammar-valid responses, and a single decision threshold is calibrated on a small development split. We compare zero-shot, instruction-guided, and few-shot prompting to strong tabular and neural baselines under identical splits, reporting accuracy, precision, recall, F1, and macro scores. Empirically, unguided prompting is unreliable, while instructions plus flags substantially improve detection quality; adding calibrated scoring further stabilizes results. On a balanced subset of two hundred flows, a 7B instruction-tuned model with flags reaches macro-F1 near 0.78; a lighter 3B model with few-shot cues and calibration attains F1 near 0.68 on one thousand examples. As the evaluation set grows to two thousand flows, decision quality decreases, revealing sensitivity to coverage and prompting. Tabular baselines remain more stable and faster, yet the prompt-only pipeline requires no gradient training, produces readable artifacts, and adapts easily through instructions and flags. Contributions include a flow-to-text protocol with interpretable cues, a calibration method for thresholding, a systematic baseline comparison, and a reproducibility bundle with prompts, grammar, metrics, and figures.

The rapid evolution of agentic AI marks a new phase in artificial intelligence, where Large Language Models (LLMs) no longer merely respond but act, reason, and adapt. This survey traces the paradigm shift in building agentic AI: from Pipeline-based systems, where planning, tool use, and memory are orchestrated by external logic, to the emerging Model-native paradigm, where these capabilities are internalized within the model's parameters. We first position Reinforcement Learning (RL) as the algorithmic engine enabling this paradigm shift. By reframing learning from imitating static data to outcome-driven exploration, RL underpins a unified solution of LLM + RL + Task across language, vision and embodied domains. Building on this, the survey systematically reviews how each capability -- Planning, Tool use, and Memory -- has evolved from externally scripted modules to end-to-end learned behaviors. Furthermore, it examines how this paradigm shift has reshaped major agent applications, specifically the Deep Research agent emphasizing long-horizon reasoning and the GUI agent emphasizing embodied interaction. We conclude by discussing the continued internalization of agentic capabilities like Multi-agent collaboration and Reflection, alongside the evolving roles of the system and model layers in future agentic AI. Together, these developments outline a coherent trajectory toward model-native agentic AI as an integrated learning and interaction framework, marking the transition from constructing systems that apply intelligence to developing models that grow intelligence through experience.

Integration of Large Language Models with search/retrieval engines has become ubiquitous, yet these systems harbor a critical vulnerability that undermines their reliability. We present the first systematic investigation of "chameleon behavior" in LLMs: their alarming tendency to shift stances when presented with contradictory questions in multi-turn conversations (especially in search-enabled LLMs). Through our novel Chameleon Benchmark Dataset, comprising 17,770 carefully crafted question-answer pairs across 1,180 multi-turn conversations spanning 12 controversial domains, we expose fundamental flaws in state-of-the-art systems. We introduce two theoretically grounded metrics: the Chameleon Score (0-1) that quantifies stance instability, and Source Re-use Rate (0-1) that measures knowledge diversity. Our rigorous evaluation of Llama-4-Maverick, GPT-4o-mini, and Gemini-2.5-Flash reveals consistent failures: all models exhibit severe chameleon behavior (scores 0.391-0.511), with GPT-4o-mini showing the worst performance. Crucially, small across-temperature variance (less than 0.004) suggests the effect is not a sampling artifact. Our analysis uncovers the mechanism: strong correlations between source re-use rate and confidence (r=0.627) and stance changes (r=0.429) are statistically significant (p less than 0.05), indicating that limited knowledge diversity makes models pathologically deferential to query framing. These findings highlight the need for comprehensive consistency evaluation before deploying LLMs in healthcare, legal, and financial systems where maintaining coherent positions across interactions is critical for reliable decision support.

As Large Language Model (LLM) agents increasingly operate in complex environments with real-world consequences, their safety becomes critical. While uncertainty quantification is well-studied for single-turn tasks, multi-turn agentic scenarios with real-world tool access present unique challenges where uncertainties and ambiguities compound, leading to severe or catastrophic risks beyond traditional text generation failures. We propose using "quitting" as a simple yet effective behavioral mechanism for LLM agents to recognize and withdraw from situations where they lack confidence. Leveraging the ToolEmu framework, we conduct a systematic evaluation of quitting behavior across 12 state-of-the-art LLMs. Our results demonstrate a highly favorable safety-helpfulness trade-off: agents prompted to quit with explicit instructions improve safety by an average of +0.39 on a 0-3 scale across all models (+0.64 for proprietary models), while maintaining a negligible average decrease of -0.03 in helpfulness. Our analysis demonstrates that simply adding explicit quit instructions proves to be a highly effective safety mechanism that can immediately be deployed in existing agent systems, and establishes quitting as an effective first-line defense mechanism for autonomous agents in high-stakes applications.

Corporate responsibility turns on notions of corporate \textit{mens rea}, traditionally imputed from human agents. Yet these assumptions are under challenge as generative AI increasingly mediates enterprise decision-making. Building on the theory of extended cognition, we argue that in response corporate knowledge may be redefined as a dynamic capability, measurable by the efficiency of its information-access procedures and the validated reliability of their outputs. We develop a formal model that captures epistemic states of corporations deploying sophisticated AI or information systems, introducing a continuous organisational knowledge metric $S_S(φ)$ which integrates a pipeline's computational cost and its statistically validated error rate. We derive a thresholded knowledge predicate $\mathsf{K}_S$ to impute knowledge and a firm-wide epistemic capacity index $\mathcal{K}_{S,t}$ to measure overall capability. We then operationally map these quantitative metrics onto the legal standards of actual knowledge, constructive knowledge, wilful blindness, and recklessness. Our work provides a pathway towards creating measurable and justiciable audit artefacts, that render the corporate mind tractable and accountable in the algorithmic age.

Robots are expected to serve as intelligent assistants, helping humans with everyday household organization. A central challenge in this setting is the task of object placement, which requires reasoning about both semantic preferences (e.g., common-sense object relations) and geometric feasibility (e.g., collision avoidance). We present GOPLA, a hierarchical framework that learns generalizable object placement from augmented human demonstrations. A multi-modal large language model translates human instructions and visual inputs into structured plans that specify pairwise object relationships. These plans are then converted into 3D affordance maps with geometric common sense by a spatial mapper, while a diffusion-based planner generates placement poses guided by test-time costs, considering multi-plan distributions and collision avoidance. To overcome data scarcity, we introduce a scalable pipeline that expands human placement demonstrations into diverse synthetic training data. Extensive experiments show that our approach improves placement success rates by 30.04 percentage points over the runner-up, evaluated on positioning accuracy and physical plausibility, demonstrating strong generalization across a wide range of real-world robotic placement scenarios.

Large language models (LLMs) have achieved remarkable success in diverse tasks, yet their safety alignment remains fragile during adaptation. Even when fine-tuning on benign data or with low-rank adaptation, pre-trained safety behaviors are easily degraded, leading to harmful responses in the fine-tuned models. To address this challenge, we propose GuardSpace, a guardrail framework for preserving safety alignment throughout fine-tuning, composed of two key components: a safety-sensitive subspace and a harmful-resistant null space. First, we explicitly decompose pre-trained weights into safety-relevant and safety-irrelevant components using covariance-preconditioned singular value decomposition, and initialize low-rank adapters from the safety-irrelevant ones, while freezing safety-relevant components to preserve their associated safety mechanism. Second, we construct a null space projector that restricts adapter updates from altering safe outputs on harmful prompts, thereby maintaining the original refusal behavior. Experiments with various pre-trained models on multiple downstream tasks demonstrate that GuardSpace achieves superior performance over existing methods. Notably, for Llama-2-7B-Chat fine-tuned on GSM8K, GuardSpace outperforms the state-of-the-art method AsFT, reducing the average harmful score from 14.4% to 3.6%, while improving the accuracy from from 26.0% to 28.0%.

The emergence of large language models (LLMs) has opened new opportunities for creating dynamic non-player characters (NPCs) in gaming environments, enabling both functional task execution and persona-consistent dialogue generation. In this paper, we (Tu_Character_lab) report our participation in the Commonsense Persona-Grounded Dialogue Challenge (CPDC) 2025 Round 2, which evaluates agents across three tracks: task-oriented dialogue, context-aware dialogue, and their integration. Our approach combines two complementary strategies: (i) lightweight prompting techniques in the API track, including a Deflanderization prompting method to suppress excessive role-play and improve task fidelity, and (ii) fine-tuned large models in the GPU track, leveraging Qwen3-14B with supervisedfinetuning (SFT) and Low-Rank Adaptation(LoRA). Our best submissions ranked 2nd on Task 1, 2nd on Task 3 (API track), and 4th on Task 3 (GPU track).