LLM-based forecasting systems have improved on real-world tasks like financial markets and sports outcomes, largely by strengthening search, but the final prediction step remains crude: systems typically ask an LLM to read all collected evidence and emit one answer. We call this approach Monolithic Prediction, which fits forecasting poorly: probabilities over competing outcomes carry information valuable for downstream decisions, but a holistic answer collapses the supporting evidence, discarding this information. To address the above issue, we propose LEAP Likelihood Elicitation and Aggregation for Probabilistic forecasting) to keep the LLM out of the final prediction. LEAP treats evidence items as local signals, uses likelihood parameters to estimate how each supports candidate outcomes, and aggregates them through a deterministic probabilistic model into a posterior over candidate outcomes, thereby preserving per-item contributions, yielding comprehensive forecasts, and simplifying audits. We build a benchmark covering real-world prediction, information-seeking, and browsing tasks, evaluating LEAP on our custom agent loop and practical agent CLI frameworks. Given the same evidence, LEAP outperforms Monolithic Prediction in most settings and metrics across models.

LLM-based multi-agent systems (MAS) are increasingly considered for high-stakes decision-making, yet outcome-based fairness audits can miss where risks arise within the decision trajectory. We present SCOPED-Hiring, a process-aware fairness diagnosis pipeline for LLM-based hiring MAS. SCOPED-Hiring constructs controlled resume variants, runs role-based hiring committees, logs over 311K structured decision trajectories, and converts trajectory fields into quantitative fairness signals organized by six diagnostic lenses: final outcome, counterfactual, process, pathway, dynamic, and design effects. SCOPED-Hiring reveals that balanced final hire rates can mask hidden trajectory unfairness in multi-agent decision trajectories: career gaps trigger suspicion, proxy cues shape qualification judgments, and identity cues lead to unequal investigation. Targeted repair guided by these diagnoses reduces total layered burden by 72.3% while shifting the hire rate by only 1.86 pp, showing that process diagnosis can guide effective repair.

Large language models (LLMs) have demonstrated multilingual capabilities, yet they are mostly English-centric due to the imbalanced training corpora. While prior works have leveraged this bias to enhance multilingual performance through translation, they have been largely limited to natural language processing (NLP) tasks. In this work, we extend the evaluation to real-world user queries and non-English-centric LLMs, offering a broader examination of multilingual performance. Our key contribution lies in demonstrating that while translation into English can boost the performance of English-centric LLMs on NLP tasks, it is not universally optimal. For culture-related tasks that need deep language understanding, prompting in the native language proves more effective as it better captures the nuances of culture and language. Our experiments expose varied behaviors across LLMs and tasks in the multilingual context, underscoring the need for a more comprehensive approach to multilingual evaluation. Therefore, we call for greater efforts in developing and evaluating LLMs that go beyond English-centric paradigms.

Large Language Models (LLMs) have shown remarkable abilities across various tasks, yet their development has predominantly centered on high-resource languages like English and Chinese, leaving low-resource languages underserved. To address this disparity, we present SeaLLMs 3, the latest iteration of the SeaLLMs model family, tailored for Southeast Asian languages. This region, characterized by its rich linguistic diversity, has lacked adequate language technology support. SeaLLMs 3 aims to bridge this gap by covering a comprehensive range of languages spoken in this region, including English, Chinese, Indonesian, Vietnamese, Thai, Tagalog, Malay, Burmese, Khmer, Lao, Tamil, and Javanese. Leveraging efficient language enhancement techniques and a specially constructed instruction tuning dataset, SeaLLMs 3 significantly reduces training costs while maintaining high performance and versatility. Our model excels in tasks such as world knowledge, mathematical reasoning, translation, and instruction following, achieving state-of-the-art performance among similarly sized models. Additionally, we prioritized safety and reliability by addressing both general and culture-specific considerations and incorporated mechanisms to reduce hallucinations. This work underscores the importance of inclusive AI, showing that advanced LLM capabilities can benefit underserved linguistic and cultural communities.

Large Language Models (LLMs) have transformed natural language processing, yet their substantial model sizes often demand significant computational resources. To preserve computing resources and accelerate inference speed, it is crucial to prune redundant parameters, especially for experienced users who often need expert models tailored to specific downstream scenarios. However, current pruning methods primarily focus on maintaining models’ general capabilities, either requiring extensive post-training or performing poorly due to coarse-grained pruning. In this work, we design a  ̲Custom  ̲Pruning method (Cus-Prun) to prune a large general model into a smaller lightweight expert model, which is positioned along the “language”, “domain” and “task” dimensions. By identifying and pruning irrelevant neurons of each dimension, Cus-Prun creates expert models without any post-training. Our experiments demonstrate that Cus-Prun consistently outperforms other methods, achieving minimal loss in both expert and general capabilities across various models from different model families and sizes.

This paper introduces a Dual Evaluation Framework to comprehensively assess the multilingual capabilities of LLMs. By decomposing the evaluation along the dimensions of linguistic medium and cultural context, this framework enables a nuanced analysis of LLMs’ ability to process questions within both native and cross-cultural contexts cross-lingually. Extensive evaluations are conducted on a wide range of models, revealing a notable “Cultural-Linguistic Synergy” phenomenon, where models exhibit better performance when questions are culturally aligned with the language. This phenomenon is further explored through interpretability probing, which shows that a higher proportion of specific neurons are activated in a language’s cultural context. This activation proportion could serve as a potential indicator for evaluating multilingual performance during model training. Our findings challenge the prevailing notion that LLMs, primarily trained on English data, perform uniformly across languages and highlight the necessity of culturally and linguistically model evaluations.

We propose a new method, Adversarial In-Context Learning (adv-ICL), to optimize prompts for in-context learning (ICL). Inspired by adversarial learning, adv-ICL is implemented as a two-player game between a generator and discriminator, with LLMs acting as both. In each round, given an input prefixed by task instructions and several exemplars, the generator produces an output. The discriminator then classifies the generator’s input-output pair as model-generated or real data. Based on the discriminator’s loss, a prompt modifier LLM proposes possible edits to the generator and discriminator prompts, and the edits that most improve the adversarial loss are selected. We show that applying adv-ICL results in significant improvements over state-of-the-art prompt optimization techniques for both open and closed-source models on 13 generation and classification tasks including summarization, arithmetic reasoning, machine translation, data-to-text generation, and the MMLU and big-bench hard benchmarks. In addition, our method is computationally efficient, easily extensible to other LLMs and tasks, and effective in low-resource settings.