We introduce the novel task of Language-Guided Object Placement in Real 3D Scenes. Our model is given a 3D scene's point cloud, a 3D asset, and a textual prompt broadly describing where the 3D asset should be placed. The task here is to find a valid placement for the 3D asset that respects the prompt. Compared with other language-guided localization tasks in 3D scenes such as grounding, this task has specific challenges: it is ambiguous because it has multiple valid solutions, and it requires reasoning about 3D geometric relationships and free space. We inaugurate this task by proposing a new benchmark and evaluation protocol. We also introduce a new dataset for training 3D LLMs on this task, as well as the first method to serve as a non-trivial baseline. We believe that this challenging task and our new benchmark could become part of the suite of benchmarks used to evaluate and compare generalist 3D LLM models.

Road traffic accidents remain a major public safety and socio-economic issue in developing countries like Bangladesh. Existing accident data collection is largely manual, fragmented, and unreliable, resulting in underreporting and inconsistent records. This research proposes a fully automated system using Large Language Models (LLMs) and web scraping techniques to address these challenges. The pipeline consists of four components: automated web scraping code generation, news collection from online sources, accident news classification with structured data extraction, and duplicate removal. The system uses the multimodal generative LLM Gemini-2.0-Flash for seamless automation. The code generation module classifies webpages into pagination, dynamic, or infinite scrolling categories and generates suitable Python scripts for scraping. LLMs also classify and extract key accident information such as date, time, location, fatalities, injuries, road type, vehicle types, and pedestrian involvement. A deduplication algorithm ensures data integrity by removing duplicate reports. The system scraped 14 major Bangladeshi news sites over 111 days (Oct 1, 2024 - Jan 20, 2025), processing over 15,000 news articles and identifying 705 unique accidents. The code generation module achieved 91.3% calibration and 80% validation accuracy. Chittagong reported the highest number of accidents (80), fatalities (70), and injuries (115), followed by Dhaka, Faridpur, Gazipur, and Cox's Bazar. Peak accident times were morning (8-9 AM), noon (12-1 PM), and evening (6-7 PM). A public repository was also developed with usage instructions. This study demonstrates the viability of an LLM-powered, scalable system for accurate, low-effort accident data collection, providing a foundation for data-driven road safety policymaking in Bangladesh.

C-to-Rust transpilation is essential for modernizing legacy C code while enhancing safety and interoperability with modern Rust ecosystems. However, no dataset currently exists for evaluating whether a system can transpile C into safe Rust that passes a set of test cases. We introduce CRUST-Bench, a dataset of 100 C repositories, each paired with manually-written interfaces in safe Rust as well as test cases that can be used to validate correctness of the transpilation. By considering entire repositories rather than isolated functions, CRUST-Bench captures the challenges of translating complex projects with dependencies across multiple files. The provided Rust interfaces provide explicit specifications that ensure adherence to idiomatic, memory-safe Rust patterns, while the accompanying test cases enforce functional correctness. We evaluate state-of-the-art large language models (LLMs) on this task and find that safe and idiomatic Rust generation is still a challenging problem for various state-of-the-art methods and techniques. We also provide insights into the errors LLMs usually make in transpiling code from C to safe Rust. The best performing model, OpenAI o1, is able to solve only 15 tasks in a single-shot setting. Improvements on CRUST-Bench would lead to improved transpilation systems that can reason about complex scenarios and help in migrating legacy codebases from C into languages like Rust that ensure memory safety. You can find the dataset and code at https://github.com/anirudhkhatry/CRUST-bench.

E-participation platforms are an important asset for governments in increasing trust and fostering democratic societies. By engaging public and private institutions and individuals, policymakers can make informed and inclusive decisions. However, current approaches of primarily static nature struggle to integrate citizen feedback effectively. Drawing on the Media Richness Theory and applying the Design Science Research method, we explore how a chatbot can address these shortcomings to improve the decision-making abilities for primary stakeholders of e-participation platforms. Leveraging the "Have Y our Say" platform, which solicits feedback on initiatives and regulations by the European Commission, a Large Language Model-based chatbot, called AskThePublic is created, providing policymakers, journalists, researchers, and interested citizens with a convenient channel to explore and engage with citizen input. Evaluating AskThePublic in 11 semi-structured interviews with public sector-affiliated experts, we find that the interviewees value the interactive and structured responses as well as enhanced language capabilities.

Effective conversational agents like large language models (LLMs) must personalize their interactions to adapt to user preferences, personalities, and attributes across diverse domains like education and healthcare. Current methods like Reinforcement Learning from Human Feedback (RLHF), often prioritize helpfulness and safety but fall short in fostering truly empathetic, adaptive, and personalized dialogues. Existing personalization approaches typically rely on extensive user history, limiting their effectiveness for new or context-limited users. To address these limitations, we propose leveraging a user model to incorporate a curiosity-based intrinsic reward into multi-turn RLHF. This novel reward mechanism encourages the LLM agent to actively infer user traits by optimizing conversations to improve its user model's accuracy. Consequently, the agent delivers more personalized interactions by learning more about the user. We demonstrate our method's effectiveness in two distinct domains: significantly improving personalization performance in a conversational recommendation task, and personalizing conversations for different learning styles in an educational setting. We show improved generalization capabilities compared to traditional multi-turn RLHF, all while maintaining conversation quality. Our method offers a promising solution for creating more personalized, adaptive, and engaging conversational agents.

Compression methods, including quantization, distillation, and pruning, improve the computational efficiency of large reasoning models (LRMs). However, existing studies either fail to sufficiently compare all three compression methods on LRMs or lack in-depth interpretation analysis. In this paper, we investigate how the reasoning capabilities of LRMs are compromised during compression, through performance benchmarking and mechanistic interpretation. To uncover the effects of compression on reasoning performance, we benchmark quantized, distilled, and pruned DeepSeek-R1 models on four reasoning datasets (AIME 2024, FOLIO, Temporal Sequences, and MuSiQue). To precisely locate compression effects on model weights, we adapt difference of means and attribution patching techniques, focusing on the activation of every linear component in compressed LRMs, to interpret fine-grained causal relationships between weights and various reasoning capabilities. This fine-grained interpretation addresses a fundamental question of compression: which weights are the most important for reasoning? Overall, we find dynamically quantized 2.51-bit R1 reaches close-to-R1 performance. With empirical verification, we present three main findings that generalize across both Llama and Qwen: (1) Weight count has a greater impact on LRMs' knowledge memorization than reasoning, highlighting the risks of pruning and distillation; (2) The MLP up projection in the final layer of distilled LRMs is one of the most important components, offering a new perspective on locating critical weights -- a fundamental problem in model compression; and (3) Current quantization methods overly compress the final-layer modules and MLP gate projections, so protecting just 2% of all weights that are excessively compressed can raise average accuracy by 6.57%, greatly surpassing the state-of-the-art. Large reasoning models (LRMs) such as DeepSeek-R1 (Guo et al., 2025) excel at complex reasoning tasks. However, due to their large sizes, deploying them can be costly and even infeasible for individuals, which hinders AI democratization.

Large Language Models (LLMs) excel at multi-step reasoning problems with explicit chain-of-thought (CoT), but verbose traces incur significant computational costs and memory overhead, and often carry redundant, stylistic artifacts. Latent reasoning has emerged as an efficient alternative that internalizes the thought process, but it suffers from a critical lack of supervision, limiting its effectiveness on complex, natural-language reasoning traces. We show that the abstract, unstructured knowledge within compressed KV -cache, which lacks direct token correspondence, can serve as a rich supervisory signal for a latent reasoning student. Empirically, the approach consistently outperforms strong latent baselines, exhibits markedly smaller degradation from equation-only to natural-language traces, and scales to larger backbones while preserving efficiency. These results establish compressed KV -cache distillation as a scalable supervision signal for latent reasoning, combining the accuracy of CoT -trained teachers with the efficiency and deployability of latent inference. Recent advancements in Large Language Models (LLMs) have demonstrated remarkable capabilities in solving complex problems across domains such as mathematics (Zhang et al., 2025), science (Phan et al., 2025), and code generation (Hui et al., 2024). A key driver of this progress is "chain-of-thought" (CoT) training that elicits intermediate steps before the final answer, improving accuracy on long-horizon inference problems (DeepSeek-AI et al., 2025). Y et, explicit CoT often incurs substantial inference cost due to long, verbose traces and the associated key-value (KV) cache growth, making deployment on memory-and compute-constrained devices difficult. Furthermore, CoT traces, especially those distilled from larger models, can inherit and amplify biases or contain plausible-sounding but fallacious logic, limiting their reliability.

We study the task of predicting dynamic physical properties from videos. More specifically, we consider physical properties that require temporal information to be inferred: elasticity of a bouncing object, viscosity of a flowing liquid, and dynamic friction of an object sliding on a surface. To this end, we make the following contributions: (i) We collect a new video dataset for each physical property, consisting of synthetic training and testing splits, as well as a real split for real world evaluation. (ii) We explore three ways to infer the physical property from videos: (a) an oracle method where we supply the visual cues that intrinsically reflect the property using classical computer vision techniques; (b) a simple read out mechanism using a visual prompt and trainable prompt vector for cross-attention on pre-trained video generative and self-supervised models; and (c) prompt strategies for Multi-modal Large Language Models (MLLMs). (iii) We show that video foundation models trained in a generative or self-supervised manner achieve a similar performance, though behind that of the oracle, and MLLMs are currently inferior to the other models, though their performance can be improved through suitable prompting.

In arena-style evaluation of large language models (LLMs), two LLMs respond to a user query, and the user chooses the winning response or deems the "battle" a draw, resulting in an adjustment to the ratings of both models. The prevailing approach for modeling these rating dynamics is to view battles as two-player game matches, as in chess, and apply the Elo rating system and its derivatives. In this paper, we critically examine this paradigm. Specifically, we question whether a draw genuinely means that the two models are equal and hence whether their ratings should be equalized. Instead, we conjecture that draws are more indicative of query difficulty: if the query is too easy, then both models are more likely to succeed equally. On three real-world arena datasets, we show that ignoring rating updates for draws yields a 1-3% relative increase in battle outcome prediction accuracy (which includes draws) for all four rating systems studied. Further analyses suggest that draws occur more for queries rated as very easy and those as highly objective, with risk ratios of 1.37 and 1.35, respectively. We recommend future rating systems to reconsider existing draw semantics and to account for query properties in rating updates.

We propose the task of knowledge distillation detection, which aims to determine whether a student model has been distilled from a given teacher, under a practical setting where only the student's weights and the teacher's API are available. This problem is motivated by growing concerns about model provenance and unauthorized replication through distillation. To address this task, we introduce a model-agnostic framework that combines data-free input synthesis and statistical score computation for detecting distillation. Our approach is applicable to both classification and generative models. Experiments on diverse architectures for image classification and text-to-image generation show that our method improves detection accuracy over the strongest baselines by 59.6% on CIFAR-10, 71.2% on ImageNet, and 20.0% for text-to-image generation. The code is available at https://github.com/shqii1j/distillation_detection.