In this work, we study the effect of annotation guidelines -- textual descriptions of event types and arguments, when instruction-tuning large language models for event extraction. We conducted a series of experiments with both human-provided and machine-generated guidelines in both full- and low-data settings. Our results demonstrate the promise of annotation guidelines when there is a decent amount of training data and highlight its effectiveness in improving cross-schema generalization and low-frequency event-type performance.

Large Language Models (LLMs) are revolutionizing information retrieval, with chatbots becoming an important source for answering user queries. As by their design, LLMs prioritize generating correct answers, the value of highly plausible yet incorrect answers (candidate answers) tends to be overlooked. However, such answers can still prove useful, for example, they can play a crucial role in tasks like Multiple-Choice Question Answering (MCQA) and QA Robustness Assessment (QARA). Existing QA datasets primarily focus on correct answers without explicit consideration of the plausibility of other candidate answers, limiting opportunity for more nuanced evaluations of models. To address this gap, we introduce PlausibleQA, a large-scale dataset comprising 10,000 questions and 100,000 candidate answers, each annotated with plausibility scores and justifications for their selection. Additionally, the dataset includes 900,000 justifications for pairwise comparisons between candidate answers, further refining plausibility assessments. We evaluate PlausibleQA through human assessments and empirical experiments, demonstrating its utility in MCQA and QARA analysis. Our findings show that plausibility-aware approaches are effective for MCQA distractor generation and QARA. We release PlausibleQA as a resource for advancing QA research and enhancing LLM performance in distinguishing plausible distractors from correct answers.

The recent application of LLMs to the legal field has spurred the creation of benchmarks across various jurisdictions and languages. However, no benchmark has yet been specifically designed for the Portuguese legal system. In this work, we present LegalBench.PT, the first comprehensive legal benchmark covering key areas of Portuguese law. To develop LegalBench.PT, we first collect long-form questions and answers from real law exams, and then use GPT-4o to convert them into multiple-choice, true/false, and matching formats. Once generated, the questions are filtered and processed to improve the quality of the dataset. To ensure accuracy and relevance, we validate our approach by having a legal professional review a sample of the generated questions. Although the questions are synthetically generated, we show that their basis in human-created exams and our rigorous filtering and processing methods applied result in a reliable benchmark for assessing LLMs' legal knowledge and reasoning abilities. Finally, we evaluate the performance of leading LLMs on LegalBench.PT and investigate potential biases in GPT-4o's responses. We also assess the performance of Portuguese lawyers on a sample of questions to establish a baseline for model comparison and validate the benchmark.

Companies across all economic sectors continue to deploy large language models at a rapid pace. Reinforcement learning is experiencing a resurgence of interest due to its association with the fine-tuning of language models from human feedback. Tool-chain language models control task-specific agents; if the converse has not already appeared, it soon will. In this paper, we present what we believe is the first investigation of an intelligent trading agent based on continuous deep reinforcement learning that also controls a large language model with which it can post to a social media feed observed by other traders. We empirically investigate the performance and impact of such an agent in a simulated financial market, finding that it learns to optimize its total reward, and thereby augment its profit, by manipulating the sentiment of the posts it produces. The paper concludes with discussion, limitations, and suggestions for future work.

The diversity of time series applications and scarcity of domain-specific data highlight the need for time-series models with strong few-shot learning capabilities. In this work, we propose a novel training scheme and a transformer-based architecture, collectively referred to as TimePFN, for multivariate time-series (MTS) forecasting. TimePFN is based on the concept of Prior-data Fitted Networks (PFN), which aims to approximate Bayesian inference. Our approach consists of (1) generating synthetic MTS data through diverse Gaussian process kernels and the linear coregionalization method, and (2) a novel MTS architecture capable of utilizing both temporal and cross-channel dependencies across all input patches. We evaluate TimePFN on several benchmark datasets and demonstrate that it outperforms the existing state-of-the-art models for MTS forecasting in both zero-shot and few-shot settings. Notably, fine-tuning TimePFN with as few as 500 data points nearly matches full dataset training error, and even 50 data points yield competitive results. We also find that TimePFN exhibits strong univariate forecasting performance, attesting to its generalization ability. Overall, this work unlocks the power of synthetic data priors for MTS forecasting and facilitates strong zero-and few-shot forecasting performance. Code -- https://github.com/egetaga/TimePFN 1 Introduction Natural language processing has achieved remarkable success driven by advances in neural architectures and data pipelines. These advances underlie modern language and vision-language models that exhibit remarkable zero-shot and few-shot learning capabilities. Inspired by these, researchers have started exploring whether such methods and ideas could be extended to time series forecasting. For instance, a notable line of work (Zhou et al. 2021; Wu et al. 2021; Zhou et al. 2022; Zhang and Y an 2023) examine the use of transformer architecture (V aswani et al. 2017) in time-series forecasting. More recently, there is also a push toward building foundation models for time series tasks (Ansari et al. 2024). However, the heterogeneous nature of time series data brings additional complications. As shown by (Zeng et al. 2023), even simple linear models are shown toTo appear in AAAI-2025 as a conference paper.

Generative AI (GenAI) tools are radically expanding the scope and capability of automation in knowledge work such as academic research. AI-assisted research tools show promise for augmenting human cognition and streamlining research processes, but could potentially increase automation bias and stifle critical thinking. We surveyed the past three years of publications from leading HCI venues. We closely examined 11 AI-assisted research tools, five employing traditional AI approaches and six integrating GenAI, to explore how these systems envision novel capabilities and design spaces. We consolidate four design recommendations that inform cognitive engagement when working with an AI research tool: Providing user agency and control; enabling divergent and convergent thinking; supporting adaptability and flexibility; and ensuring transparency and accuracy. We discuss how these ideas mark a shift in AI-assisted research tools from mimicking a researcher's established workflows to generative co-creation with the researcher and the opportunities this shift affords the research community.

Generative AI (GenAI) is increasingly used in survey contexts to simulate human preferences. While many research endeavors evaluate the quality of synthetic GenAI data by comparing model-generated responses to gold-standard survey results, fundamental questions about the validity and reliability of using LLMs as substitutes for human respondents remain. Our study provides a technical analysis of how demographic attributes and prompt variations influence latent opinion mappings in large language models (LLMs) and evaluates their suitability for survey-based predictions. Using 14 different models, we find that LLM-generated data fails to replicate the variance observed in real-world human responses, particularly across demographic subgroups. In the political space, persona-to-party mappings exhibit limited differentiation, resulting in synthetic data that lacks the nuanced distribution of opinions found in survey data. Moreover, we show that prompt sensitivity can significantly alter outputs for some models, further undermining the stability and predictiveness of LLM-based simulations. As a key contribution, we adapt a probe-based methodology that reveals how LLMs encode political affiliations in their latent space, exposing the systematic distortions introduced by these models. Our findings highlight critical limitations in AI-generated survey data, urging caution in its use for public opinion research, social science experimentation, and computational behavioral modeling.

UMBC CODEBOT '25 Workshop Columbia, MD / 25-26 February 2025BEYOND TRUSTING TRUST: MUL TI-MODEL V ALIDA TION FOR ROBUST CODE GENERA TION Bradley McDanel Franklin and Marshall College bmcdanel@fandm.edu 1 Introduction Ken Thompson's 1984 essay "Reflections on Trusting Trust" demonstrated that even carefully reviewed source code could hide malicious behavior through compromised compilers - because the malicious code exists only in the compiled binary form, not its source [1]. Today, large language models (LLMs) used as code generators [2, 3] present an even more opaque security challenge than classical compilers. While compiler binaries can be analyzed for malicious behavior, LLMs operate through vast matrices of weights combined in non-linear ways, making it difficult to develop robust methods for identifying embedded behaviors [4, 5]. This paper revisits Thompson's analogy in the context of LLM-based code generation. We show how malicious behavior might be subtly embedded into a widely used model and argue that direct inspection of the model's parameters is currently infeasible.

--The Qwen 2.5 3B base model was fine-tuned to generate contextually rich and engaging movie dialogue, leveraging the Cornell Movie-Dialog Corpus, a curated dataset of movie conversations. Due to limitations in GPU computing and VRAM, the training process began with the 0.5B model, progressively scaling up to the 1.5B and 3B versions as efficiency improvements were implemented. The Qwen 2.5 series, developed by Alibaba Group, stands at the forefront of small open-source pre-trained models, particularly excelling in creative tasks compared to alternatives like Meta's Llama 3.2 and Google's Gemma. Results demonstrate the ability of small models to produce high-quality, realistic dialogue, offering a promising approach for real-time, context-sensitive conversation generation. This project aimed to fine-tune a small, pretrained large language model (LLM) to generate realistic and compelling movie dialogue when prompted with a preceding line. To achieve this, the Qwen 2.5 3B base model was fine-tuned using the Cornell Movie-Dialog Corpus, a curated dataset of movie dialogue [1].

Evaluating large language models (LLMs) poses significant challenges, particularly due to issues of data contamination and the leakage of correct answers. To address these challenges, we introduce ThinkBench, a novel evaluation framework designed to evaluate LLMs' reasoning capability robustly. ThinkBench proposes a dynamic data generation method for constructing out-of-distribution (OOD) datasets and offers an OOD dataset that contains 2,912 samples drawn from reasoning tasks. ThinkBench unifies the evaluation of reasoning models and non-reasoning models. We evaluate 16 LLMs and 4 PRMs under identical experimental conditions and show that most of the LLMs' performance are far from robust and they face a certain level of data leakage. By dynamically generating OOD datasets, ThinkBench effectively provides a reliable evaluation of LLMs and reduces the impact of data contamination.