Spatiotemporal relationships are critical in data science, as many prediction and reasoning tasks require analysis across both spatial and temporal dimensions--for instance, navigating an unfamiliar city involves planning itineraries that sequence locations and timing cultural experiences. However, existing Question-Answering (QA) datasets lack sufficient spatiotemporal-sensitive questions, making them inadequate benchmarks for evaluating models' spatiotemporal reasoning capabilities. To address this gap, we introduce POI-QA, a novel spatiotemporal-sensitive QA dataset centered on Point of Interest (POI), constructed through three key steps: mining and aligning open-source vehicle trajectory data from GAIA with high-precision geographic POI data, rigorous manual validation of noisy spatiotemporal facts, and generating bilingual (Chinese/English) QA pairs that reflect human-understandable spatiotemporal reasoning tasks. Our dataset challenges models to parse complex spatiotemporal dependencies, and evaluations of state-of-the-art multilingual LLMs (e.g., Qwen2.5-7B, Llama3.1-8B) reveal stark limitations: even the top-performing model (Qwen2.5-7B fine-tuned with RAG+LoRA) achieves a top 10 Hit Ratio (HR@10) of only 0.41 on the easiest task, far below human performance at 0.56. This underscores persistent weaknesses in LLMs' ability to perform consistent spatiotemporal reasoning, while highlighting POI-QA as a robust benchmark to advance algorithms sensitive to spatiotemporal dynamics. The dataset is publicly available at https://www.kaggle.com/ds/7394666.

Theory of Mind (ToM), the ability to attribute mental states to others, is fundamental for human social intelligence and a critical capability for advanced Artificial Intelligence. Recent advancements in Large Language Models (LLMs) have shown promising performance on ToM benchmarks, raising the question: Do these benchmarks necessitate explicit human-like reasoning processes, or can models succeed through alternative strategies? We investigate this question empirically by applying Reinforcement Learning (RL) and Supervised Fine-Tuning (SFT) to LLMs of varying scales (0.5B to 7B parameters) and evaluating them across multiple ToM datasets. Our results reveal a scale-dependent impact of RL: while RL significantly improves accuracy and fosters high-quality, interpretable, and transferable belief-tracking reasoning in larger models (7B), it leads to "reasoning collapse" in smaller models ($\leq$3B), where high accuracy and generalization ability are achieved via drastically shortened, less meaningful responses. Surprisingly, further SFT achieves competitive and generalizable performance across these benchmarks, often matching or exceeding RL models in accuracy, despite not being explicitly trained to produce structured reasoning traces. These findings highlight a critical discrepancy between benchmark accuracy and the nature of learned reasoning. Our work suggests that current ToM benchmarks may be solvable without requiring the explicit, human-like simulation of mental states they were designed to probe. LLMs, particularly when scale is limited or training signals focus solely on output correctness, may leverage alternative rules effective for benchmark data structures.

Patent claims define the scope of protection and establish the legal boundaries of an invention. Drafting these claims is a complex and time-consuming process that usually requires the expertise of skilled patent attorneys, which can form a large access barrier for many small enterprises. To solve these challenges, researchers have investigated the use of large language models (LLMs) for automating patent claim generation. However, existing studies highlight inconsistencies between automated evaluation metrics and human expert assessments. To bridge this gap, we introduce Patent-CE, the first comprehensive benchmark for evaluating patent claims. Patent-CE includes comparative claim evaluations annotated by patent experts, focusing on five key criteria: feature completeness, conceptual clarity, terminology consistency, logical linkage, and overall quality. Additionally, we propose PatClaimEval, a novel multi-dimensional evaluation method specifically designed for patent claims. Our experiments demonstrate that PatClaimEval achieves the highest correlation with human expert evaluations across all assessment criteria among all tested metrics. This research provides the groundwork for more accurate evaluations of automated patent claim generation systems.

Multimodal Large Language Models which can answer complex questions on an image struggle to tell the time on analog clocks. This is probably due to the lack of images with clocks at different times in their training set. In this work we explore this issue with one of the latest MLLMs: GPT-4.1 to understand why MLLMs fail to tell the time and whether fine-tuning can solve the problem. The results show how models are making progress in reading the time on analog clocks. But have they really learned to do it, or have they only learned patterns in their training datasets? In this work we put the models to the test with different clocks to illustrate the limitations of MLLMs to abstract and generalize.

Verifying the authenticity of AI-generated images presents a growing challenge on social media platforms these days. While vision-language models (VLMs) like CLIP outdo in multimodal representation, their capacity for AI-generated image classification is underexplored due to the absence of such labels during the pre-training process. This work investigates whether CLIP embeddings inherently contain information indicative of AI generation. A proposed pipeline extracts visual embeddings using a frozen CLIP model, feeds its embeddings to lightweight networks, and fine-tunes only the final classifier. Experiments on the public CIFAKE benchmark show the performance reaches 95% accuracy without language reasoning. Few-shot adaptation to curated custom with 20% of the data results in performance to 85%. A closed-source baseline (Gemini-2.0) has the best zero-shot accuracy yet fails on specific styles. Notably, some specific image types, such as wide-angle photographs and oil paintings, pose significant challenges to classification. These results indicate previously unexplored difficulties in classifying certain types of AI-generated images, revealing new and more specific questions in this domain that are worth further investigation.

We investigate the usage of Large Language Model (LLM) in collecting high-quality data to warm-start Reinforcement Learning (RL) algorithms for learning in some classical Markov Decision Process (MDP) environments. In this work, we focus on using LLM to generate an off-policy dataset that sufficiently covers state-actions visited by optimal policies, then later using an RL algorithm to explore the environment and improve the policy suggested by the LLM. Our algorithm, LORO, can both converge to an optimal policy and have a high sample efficiency thanks to the LLM's good starting policy. On multiple OpenAI Gym environments, such as CartPole and Pendulum, we empirically demonstrate that LORO outperforms baseline algorithms such as pure LLM-based policies, pure RL, and a naive combination of the two, achieving up to $4 \times$ the cumulative rewards of the pure RL baseline.

This paper systematically addresses the challenges of rule retrieval, a crucial yet underexplored area. Vanilla retrieval methods using sparse or dense retrievers to directly search for relevant rules to support downstream reasoning, often suffer from low accuracy. This is primarily due to a significant semantic gap between the instantiated facts in the queries and the abstract representations of the rules. Such misalignment results in suboptimal retrieval quality, which in turn negatively impacts reasoning performance. To overcome these challenges, we propose Self-Induction Augmented Retrieval (SIAR), a novel approach that utilizes Large Language Models (LLMs) to induce potential inferential rules that might offer benefits for reasoning by abstracting the underlying knowledge and logical structure in queries. These induced rules are then used for query augmentation to improve retrieval effectiveness. Additionally, we introduce Rule Relevance ReEstimate (R$^3$), a method that re-estimates the relevance of retrieved rules by assessing whether the abstract knowledge they contain can be instantiated to align with the facts in the queries and the helpfulness for reasoning. Extensive experiments across various settings demonstrate the effectiveness and versatility of our proposed methods.

Large Language Models (LLMs) excel in various domains but pose inherent privacy risks. Existing methods to evaluate privacy leakage in LLMs often use memorized prefixes or simple instructions to extract data, both of which well-alignment models can easily block. Meanwhile, Jailbreak attacks bypass LLM safety mechanisms to generate harmful content, but their role in privacy scenarios remains underexplored. In this paper, we examine the effectiveness of jailbreak attacks in extracting sensitive information, bridging privacy leakage and jailbreak attacks in LLMs. Moreover, we propose PIG, a novel framework targeting Personally Identifiable Information (PII) and addressing the limitations of current jailbreak methods. Specifically, PIG identifies PII entities and their types in privacy queries, uses in-context learning to build a privacy context, and iteratively updates it with three gradient-based strategies to elicit target PII. We evaluate PIG and existing jailbreak methods using two privacy-related datasets. Experiments on four white-box and two black-box LLMs show that PIG outperforms baseline methods and achieves state-of-the-art (SoTA) results. The results underscore significant privacy risks in LLMs, emphasizing the need for stronger safeguards. Our code is availble at https://github.com/redwyd/PrivacyJailbreak.

With the abundance of information in today's digital world, it is a major challenge to process voluminous text from news articles, reports, and web pages in an efficient manner. Text summarization solves this problem by providing brief, informative summaries of lengthy documents, both saving end-users time and mental effort [1]. Whereas traditional summarization methods involve only extractive approaches (identifying major sentences out of the source text) and abstractive approaches (producing new sentences capturing the core meaning), the current project outlines a holistic, multi-step NLP pipeline extending beyond mere summarization efforts [1]. The pipeline starts with Optical Character Recognition (OCR), which is achieved with Tesseract (Pytesseract). This module yields machine-readable text from images and handles various languages such as English, Hindi, Tamil, Urdu, Bengali, and Telugu [1]. The extracted information then passes through a chain of Natural Language Processing (NLP) and Machine Learning (ML) modules for more in-depth text analysis. The main elements of this pipeline are: The system combines state-of-the-art NLP features to boost text comprehension and processing.

Time series forecasting is critical across multiple domains, where time series data exhibits both local patterns and global dependencies. While Transformer-based methods effectively capture global dependencies, they often overlook short-term local variations in time series. Recent methods that adapt large language models (LLMs) into time series forecasting inherit this limitation by treating LLMs as black-box encoders, relying solely on the final-layer output and underutilizing hierarchical representations. To address this limitation, we propose Logo-LLM, a novel LLM-based framework that explicitly extracts and models multi-scale temporal features from different layers of a pre-trained LLM. Through empirical analysis, we show that shallow layers of LLMs capture local dynamics in time series, while deeper layers encode global trends. Moreover, Logo-LLM introduces lightweight Local-Mixer and Global-Mixer modules to align and integrate features with the temporal input across layers. Extensive experiments demonstrate that Logo-LLM achieves superior performance across diverse benchmarks, with strong generalization in few-shot and zero-shot settings while maintaining low computational overhead.