We propose GuessBench, a novel benchmark that evaluates Vision Language Models (VLMs) on modeling the pervasive, noisy, and pluralistic human creativity. GuessBench sources data from "Guess the Build", an online multiplayer Minecraft minigame where one player constructs a Minecraft build given a concept (e.g. caterpillar) and others try to guess it with natural language hints, presenting a pristine testbed for sensemaking creativity in the wild with VLMs acting as guessers. We curate 1500 images from the actual gameplay and design 2000 problems spanning static and dynamic image settings, natural language hints of varying completeness, and more. Extensive experiments with six open/API VLMs and five reasoning enhancement approaches demonstrate that GuessBench presents a uniquely challenging task in creativity modeling: even the start-of-the-art GPT-4o is incorrect on 34% of instances, while we observe a huge performance gap (13.87% vs. 53.93% on average) between open and API models. When used as a resource to improve VLMs, fine-tuning on the reasoning traces for GuessBench problems improves visual perception tasks by 15.36% on average. Further analysis reveals that VLM performance in creativity sensemaking correlates with the frequency of the concept in training data, while the accuracy drops sharply for concepts in underrepresented cultural contexts and low-resource languages.

We propose Data Swarms, an algorithm to optimize the generation of synthetic evaluation data and advance quantitative desiderata of LLM evaluation. We first train a swarm of initial data generators using existing data, and define various evaluation objectives to reflect the desired properties of evaluation (e.g., generate more difficult problems for the evaluated models) and quantitatively evaluate data generators. We then employ particle swarm optimization to optimize the swarm of data generators, where they collaboratively search through the model parameter space to find new generators that advance these objectives. We further extend it to Adversarial Swarms, where the data generator swarm generates harder data while the test taker model swarm learns from such data, co-evolving dynamically for better data and models simultaneously. Extensive experiments demonstrate that Data Swarms outperforms eight data generation baselines across five evaluation objectives, while Adversarial Swarms produce more robust learning of synthetic data and stronger generalization. Further analysis reveals that Data Swarms successfully optimizes compositions of multiple evaluation objectives and generalizes to new off-the-shelf LLMs, unseen at optimization time.

The scientific literature is growing rapidly, making it hard to keep track of the state-of-the-art. Systematic literature reviews (SLRs) aim to identify and evaluate all relevant papers on a topic. After retrieving a set of candidate papers, the abstract screening phase determines initial relevance. To date, abstract screening methods using large language models (LLMs) focus on binary classification settings; existing question answering (QA) based ranking approaches suffer from error propagation. LLMs offer a unique opportunity to evaluate the SLR's inclusion and exclusion criteria, yet, existing benchmarks do not provide them exhaustively. We manually extract these criteria as well as research questions for 57 SLRs, mostly in the medical domain, enabling principled comparisons between approaches. Moreover, we propose LGAR, a zero-shot LLM Guided Abstract Ranker composed of an LLM based graded relevance scorer and a dense re-ranker. Our extensive experiments show that LGAR outperforms existing QA-based methods by 5-10 pp. in mean average precision. Our code and data is publicly available.

Graph-based RAG methods like GraphRAG have shown promising global understanding of the knowledge base by constructing hierarchical entity graphs. However, they often suffer from inefficiency and rely on manually pre-defined query modes, limiting practical use. In this paper, we propose E^2GraphRAG, a streamlined graph-based RAG framework that improves both Efficiency and Effectiveness. During the indexing stage, E^2GraphRAG constructs a summary tree with large language models and an entity graph with SpaCy based on document chunks. We then construct bidirectional indexes between entities and chunks to capture their many-to-many relationships, enabling fast lookup during both local and global retrieval. For the retrieval stage, we design an adaptive retrieval strategy that leverages the graph structure to retrieve and select between local and global modes. Experiments show that E^2GraphRAG achieves up to 10 times faster indexing than GraphRAG and 100 times speedup over LightRAG in retrieval while maintaining competitive QA performance.

Large Language Models (LLMs) show promise as data analysis agents, but existing benchmarks overlook the iterative nature of the field, where experts' decisions evolve with deeper insights of the dataset. To address this, we introduce IDA-Bench, a novel benchmark evaluating LLM agents in multi-round interactive scenarios. Derived from complex Kaggle notebooks, tasks are presented as sequential natural language instructions by an LLM-simulated user. Agent performance is judged by comparing its final numerical output to the human-derived baseline. Initial results show that even state-of-the-art coding agents (like Claude-3.7-thinking) succeed on < 50% of the tasks, highlighting limitations not evident in single-turn tests. This work underscores the need to improve LLMs' multi-round capabilities for building more reliable data analysis agents, highlighting the necessity of achieving a balance between instruction following and reasoning.

The 5W1H questions -- who, what, when, where, why and how -- are commonly used in journalism to ensure that an article describes events clearly and systematically. Answering them is a crucial prerequisites for tasks such as summarization, clustering, and news aggregation. In this paper, we design the first automated extraction pipeline to get 5W1H information from French news articles. To evaluate the performance of our algorithm, we also create a corpus of 250 Quebec news articles with 5W1H answers marked by four human annotators. Our results demonstrate that our pipeline performs as well in this task as the large language model GPT-4o.

The protection of cyber Intellectual Property (IP) such as web content is an increasingly critical concern. The rise of large language models (LLMs) with online retrieval capabilities enables convenient access to information but often undermines the rights of original content creators. As users increasingly rely on LLM-generated responses, they gradually diminish direct engagement with original information sources, which will significantly reduce the incentives for IP creators to contribute, and lead to a saturating cyberspace with more AI-generated content. In response, we propose a novel defense framework that empowers web content creators to safeguard their web-based IP from unauthorized LLM real-time extraction and redistribution by leveraging the semantic understanding capability of LLMs themselves. Our method follows principled motivations and effectively addresses an intractable black-box optimization problem. Real-world experiments demonstrated that our methods improve defense success rates from 2.5% to 88.6% on different LLMs, outperforming traditional defenses such as configuration-based restrictions.

Recent advances in large-scale generative language models have shown that reasoning capabilities can significantly improve model performance across a variety of tasks. However, the impact of reasoning on a model's ability to mitigate stereotypical responses remains largely underexplored. In this work, we investigate the crucial relationship between a model's reasoning ability and fairness, and ask whether improved reasoning capabilities can mitigate harmful stereotypical responses, especially those arising due to shallow or flawed reasoning. We conduct a comprehensive evaluation of multiple open-source LLMs, and find that larger models with stronger reasoning abilities exhibit substantially lower stereotypical bias on existing fairness benchmarks. Building on this insight, we introduce ReGiFT -- Reasoning Guided Fine-Tuning, a novel approach that extracts structured reasoning traces from advanced reasoning models and infuses them into models that lack such capabilities. We use only general-purpose reasoning and do not require any fairness-specific supervision for bias mitigation. Notably, we see that models fine-tuned using ReGiFT not only improve fairness relative to their non-reasoning counterparts but also outperform advanced reasoning models on fairness benchmarks. We also analyze how variations in the correctness of the reasoning traces and their length influence model fairness and their overall performance. Our findings highlight that enhancing reasoning capabilities is an effective, fairness-agnostic strategy for mitigating stereotypical bias caused by reasoning flaws.

Large Language Models (LLMs) trained on historical web data inevitably become outdated. We investigate evaluation strategies and update methods for LLMs as new data becomes available. We introduce a web-scale dataset for time-continual pretraining of LLMs derived from 114 dumps of Common Crawl (CC) - orders of magnitude larger than previous continual language modeling benchmarks. We also design time-stratified evaluations across both general CC data and specific domains (Wikipedia, StackExchange, and code documentation) to assess how well various continual learning methods adapt to new data while retaining past knowledge. Our findings demonstrate that, on general CC data, autoregressive meta-schedules combined with a fixed-ratio replay of older data can achieve comparable held-out loss to re-training from scratch, while requiring significantly less computation (2.6x). However, the optimal balance between incorporating new data and replaying old data differs as replay is crucial to avoid forgetting on generic web data but less so on specific domains.

The US administration's targeting of academic research and international students is a "great gift" to China in the race to compete on artificial intelligence, former OpenAI board member Helen Toner has said. The director of strategy at Georgetown's Center for Security and Emerging Technology (CSET) joined the board of OpenAI in 2021 after a career studying AI and the relationship between the United States and China. Toner, a 33-year-old University of Melbourne graduate, was on the board for two years until a falling out with founder Sam Altman in 2023. Altman was fired by the board over claims that he was not "consistently candid" in his communications and the board did not have confidence in Altman's ability to lead. The chaotic months that followed saw Altman fired and then re-hired with three members of the board, including Toner, ousted instead.