We present a comprehensive evaluation of gender fairness in large language models (LLMs), focusing on their ability to handle both binary and non-binary genders. While previous studies primarily focus on binary gender distinctions, we introduce the Gender Inclusivity Fairness Index (GIFI), a novel and comprehensive metric that quantifies the diverse gender inclusivity of LLMs. GIFI consists of a wide range of evaluations at different levels, from simply probing the model with respect to provided gender pronouns to testing various aspects of model generation and cognitive behaviors under different gender assumptions, revealing biases associated with varying gender identifiers. We conduct extensive evaluations with GIFI on 22 prominent open-source and proprietary LLMs of varying sizes and capabilities, discovering significant variations in LLMs’ gender inclusivity. Our study highlights the importance of improving LLMs’ inclusivity, providing a critical benchmark for future advancements in gender fairness in generative models.

Conditional dichotomy, the contrast between two outputs conditioned on the same context, is vital for applications such as debate, defeasible inference, and causal reasoning. Existing methods that rely on semantic similarity often fail to capture the nuanced oppositional dynamics essential for these applications. Motivated by these limitations, we introduce a novel task, Conditional Dichotomy Quantification (ConDQ), which formalizes the direct measurement of conditional dichotomy and provides carefully constructed datasets covering debate, defeasible natural language inference, and causal reasoning scenarios. To address this task, we develop the Dichotomy-oriented Geometric Embedding (DoGE) framework, which leverages complex-valued embeddings and a dichotomous objective to model and quantify these oppositional relationships effectively. Extensive experiments validate the effectiveness and versatility of DoGE, demonstrating its potential in understanding and quantifying conditional dichotomy across diverse NLP applications. Our code and datasets are available at https://github.com/cui-shaobo/conditional-dichotomy-quantification.

Process-level Reward Models (PRMs) are crucial for complex reasoning and decision-making tasks, where each intermediate step plays an important role in the reasoning process. Since language models are prone to various types of errors during the reasoning process, PRMs are required to possess nuanced capabilities for detecting various implicit error types in real-world scenarios. However, current benchmarks primarily focus on step correctness, failing to evaluate PRMs’ performance systematically. To address this gap, we introduce PRMBench, a process-level benchmark specifically designed to assess the fine-grained error detection capabilities of PRMs. PRMBench comprises 6,216 carefully designed problems and 83,456 step-level labels, evaluating models across multiple dimensions, including simplicity, soundness, and sensitivity. In our experiments on 25 models, spanning both open-source PRMs and closed-source large language models prompted as critic models, we uncover significant weaknesses in current PRMs. These findings underscore the challenges inherent in process-level evaluation and highlight key directions for future research, establishing PRMBench as a robust testbed for advancing research on PRM evaluation and development.

We introduce VLM-Lens, a toolkit designed to enable systematic benchmarking, analysis, and interpretation of vision-language models (VLMs) by supporting the extraction of intermediate outputs from any layer during the forward pass of open-source VLMs. VLM-Lens provides a unified, YAML-configurable interface that abstracts away model-specific complexities and supports user-friendly operation across diverse VLMs. It currently supports 16 state-of-the-art base VLMs and their over 30 variants, and is extensible to accommodate new models without changing the core logic.The toolkit integrates easily with various interpretability and analysis methods. We demonstrate its usage with two simple analytical experiments, revealing systematic differences in the hidden representations of VLMs across layers and target concepts. VLM-Lens is released as an open-sourced project to accelerate community efforts in understanding and improving VLMs.

As large language models (LLMs) become increasingly integrated into personal writing tools, a critical question arises: can LLMs faithfully imitate an individual’s writing style from just a few examples? Personal style is often subtle and implicit, making it difficult to specify through prompts yet essential for user-aligned generation. This work presents a comprehensive evaluation of state-of-the-art LLMs’ ability to mimic personal writing styles via in-context learning from a small number of user-authored samples. We introduce an ensemble of complementary metrics—including authorship attribution, authorship verification, style matching, and AI detection—to robustly assess style imitation. Our evaluation spans over 40,000 generations per model across domains such as news, email, forums, and blogs, covering writing samples from more than 400 real-world authors. Results show that while LLMs can approximate user styles in structured formats like news and email, they struggle with nuanced, informal writing in blogs and forums. Further analysis on various prompting strategies such as number of demonstrations reveal key limitations in effective personalization. Our findings highlight a fundamental gap in personalized LLM adaptation and the need for improved techniques to support implicit, style-consistent generation. To aid future research and for reproducibility, we open-source our data and code.

Large language models (LLMs) and their multimodal variants can now process visual inputs, including images of text. This raises an intriguing question: Can we compress textual inputs by feeding them as images to reduce token usage while preserving performance?In this paper, we show that *visual text representations* are a practical and surprisingly effective form of input compression for decoder LLMs. We exploit this idea by rendering long text inputs as a single image and providing it directly to the model. This approach dramatically reduces the number of decoder tokens required, offering a new form of input compression. Through experiments on two distinct benchmarks — RULER (long-context retrieval) and CNN/DailyMail (document summarization) — we demonstrate that this text-as-image method yields substantial token savings *without degrading task performance*.

Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning, positioning them as promising tools for supporting human problem-solving. However, what happens when their performance is affected by *misinformation*, i.e., incorrect inputs introduced by users due to oversights or gaps in knowledge? Such misinformation is prevalent in real-world interactions with LLMs, yet how it propagates within LLMs’ reasoning process remains underexplored. Focusing on mathematical reasoning, we present a comprehensive analysis of how misinformation affects intermediate reasoning steps and final answers. We also examine how effectively LLMs can correct misinformation when explicitly instructed to do so. Even with explicit instructions, LLMs succeed less than half the time in rectifyingmisinformation, despite possessing correct internal knowledge, leading to significant accuracy drops (10.02% – 72.20%), and the degradation holds with thinking models (4.30% – 19.97%). Further analysis shows that applying factual corrections early in the reasoning process most effectively reduces misinformation propagation, and fine-tuning on synthesized data with early-stage corrections significantly improves reasoning factuality. Our work offers a practical approach to mitigating misinformation propagation.

Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models’ ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available [here](https://github.com/jeremyxianx/Assisted-DS).

There has recently been considerable interest in incorporating information retrieval into large language models (LLMs). Retrieval from a dynamically expanding external corpus of text allows a model to incorporate current events and can be viewed as a form of episodic memory. Here we demonstrate that pre-processing the external corpus into semi-structured “atomic facts” makes retrieval more efficient. More specifically, we demonstrate that our particular form of atomic facts improves performance on various question answering tasks when the amount of retrieved text is limited. Limiting the amount of retrieval reduces the size of the context and improves inference efficiency.

Violence-provoking speech – speech that implicitly or explicitly promotes violence against the members of the targeted community, contributed to a massive surge in anti-Asian crimes during the COVID-19 pandemic. While previous works have characterized and built tools for detecting other forms of harmful speech, like fear speech and hate speech, our work takes a community-centric approach to studying anti-Asian violence-provoking speech. Using data from ~420k Twitter posts spanning a 3-year duration (January 1, 2020 to February 1, 2023), we develop a codebook to characterize anti-Asian violence-provoking speech and collect a community-crowdsourced dataset to facilitate its large-scale detection using state-of-the-art classifiers. We contrast the capabilities of natural language processing classifiers, ranging from BERT-based to LLM-based classifiers, in detecting violence-provoking speech with their capabilities to detect anti-Asian hateful speech. In contrast to prior work that has demonstrated the effectiveness of such classifiers in detecting hateful speech (F₁ = 0.89), our work shows that accurate and reliable detection of violence-provoking speech is a challenging task (F₁ = 0.69). We discuss the implications of our findings, particularly the need for proactive interventions to support Asian communities during public health crises.

The field of unsupervised machine translation has seen significant advancement from the marriage of the Transformer and the back-translation algorithm. The Transformer is a powerful generative model, and back-translation leverages Transformer’s high-quality translations for iterative self-improvement. However, the Transformer is encumbered by the run-time of autoregressive inference during back-translation, and back-translation is limited by a lack of synthetic data efficiency. We propose a two-for-one improvement to Transformer back-translation: Quick Back-Translation (QBT). QBT re-purposes the encoder as a generative model, and uses encoder-generated sequences to train the decoder in conjunction with the original autoregressive back-translation step, improving data throughput and utilization. Experiments on various WMT benchmarks demonstrate that a relatively small number of refining steps of QBT improve current unsupervised machine translation models, and that QBT dramatically outperforms standard back-translation only method in terms of training efficiency for comparable translation qualities.

Standard conversational semantic parsing maps a complete user utterance into an executable program, after which the program is executed to respond to the user. This could be slow when the program contains expensive function calls. We investigate the opportunity to reduce latency by predicting and executing function calls while the user is still speaking. We introduce the task of online semantic parsing for this purpose, with a formal latency reduction metric inspired by simultaneous machine translation. We propose a general framework with first a learned prefix-to-program prediction module, and then a simple yet effective thresholding heuristic for subprogram selection for early execution. Experiments on the SMCalFlow and TreeDST datasets show our approach achieves large latency reduction with good parsing quality, with a 30%–65% latency reduction depending on function execution time and allowed cost.

To alleviate the data scarcity problem in training question answering systems, recent works propose additional intermediate pre-training for dense passage retrieval (DPR). However, there still remains a large discrepancy between the provided upstream signals and the downstream question-passage relevance, which leads to less improvement. To bridge this gap, we propose the HyperLink-induced Pre-training (HLP), a method to pre-train the dense retriever with the text relevance induced by hyperlink-based topology within Web documents. We demonstrate that the hyperlink-based structures of dual-link and co-mention can provide effective relevance signals for large-scale pre-training that better facilitate downstream passage retrieval. We investigate the effectiveness of our approach across a wide range of open-domain QA datasets under zero-shot, few-shot, multi-hop, and out-of-domain scenarios. The experiments show our HLP outperforms the BM25 by up to 7 points as well as other pre-training methods by more than 10 points in terms of top-20 retrieval accuracy under the zero-shot scenario. Furthermore, HLP significantly outperforms other pre-training methods under the other scenarios.

Evaluations in machine learning rarely use the latest metrics, datasets, or human evaluation in favor of remaining compatible with prior work. The compatibility, often facilitated through leaderboards, thus leads to outdated but standardized evaluation practices. We pose that the standardization is taking place in the wrong spot. Evaluation infrastructure should enable researchers to use the latest methods and what should be standardized instead is how to incorporate these new evaluation advances. We introduce GEMv2, the new version of the Generation, Evaluation, and Metrics Benchmark which uses a modular infrastructure for dataset, model, and metric developers to benefit from each other’s work. GEMv2 supports 40 documented datasets in 51 languages, ongoing online evaluation for all datasets, and our interactive tools make it easier to add new datasets to the living benchmark.

Transition-based parsers for Abstract Meaning Representation (AMR) rely on node-to-word alignments. These alignments are learned separately from parser training and require a complex pipeline of rule-based components, pre-processing, and post-processing to satisfy domain-specific constraints. Parsers also train on a point-estimate of the alignment pipeline, neglecting the uncertainty due to the inherent ambiguity of alignment. In this work we explore two avenues for overcoming these limitations. First, we propose a neural aligner for AMR that learns node-to-word alignments without relying on complex pipelines. We subsequently explore a tighter integration of aligner and parser training by considering a distribution over oracle action sequences arising from aligner uncertainty. Empirical results show this approach leads to more accurate alignments and generalization better from the AMR2.0 to AMR3.0 corpora. We attain a new state-of-the art for gold-only trained models, matching silver-trained performance without the need for beam search on AMR3.0.

Predicting linearized Abstract Meaning Representation (AMR) graphs using pre-trained sequence-to-sequence Transformer models has recently led to large improvements on AMR parsing benchmarks. These parsers are simple and avoid explicit modeling of structure but lack desirable properties such as graph well-formedness guarantees or built-in graph-sentence alignments. In this work we explore the integration of general pre-trained sequence-to-sequence language models and a structure-aware transition-based approach. We depart from a pointer-based transition system and propose a simplified transition set, designed to better exploit pre-trained language models for structured fine-tuning. We also explore modeling the parser state within the pre-trained encoder-decoder architecture and different vocabulary strategies for the same purpose. We provide a detailed comparison with recent progress in AMR parsing and show that the proposed parser retains the desirable properties of previous transition-based approaches, while being simpler and reaching the new parsing state of the art for AMR 2.0, without the need for graph re-categorization.

We introduce GEM, a living benchmark for natural language Generation (NLG), its Evaluation, and Metrics. Measuring progress in NLG relies on a constantly evolving ecosystem of automated metrics, datasets, and human evaluation standards. Due to this moving target, new models often still evaluate on divergent anglo-centric corpora with well-established, but flawed, metrics. This disconnect makes it challenging to identify the limitations of current models and opportunities for progress. Addressing this limitation, GEM provides an environment in which models can easily be applied to a wide set of tasks and in which evaluation strategies can be tested. Regular updates to the benchmark will help NLG research become more multilingual and evolve the challenge alongside models. This paper serves as the description of the data for the 2021 shared task at the associated GEM Workshop.

Abstract Meaning Representation parsing is a sentence-to-graph prediction task where target nodes are not explicitly aligned to sentence tokens. However, since graph nodes are semantically based on one or more sentence tokens, implicit alignments can be derived. Transition-based parsers operate over the sentence from left to right, capturing this inductive bias via alignments at the cost of limited expressiveness. In this work, we propose a transition-based system that combines hard-attention over sentences with a target-side action pointer mechanism to decouple source tokens from node representations and address alignments. We model the transitions as well as the pointer mechanism through straightforward modifications within a single Transformer architecture. Parser state and graph structure information are efficiently encoded using attention heads. We show that our action-pointer approach leads to increased expressiveness and attains large gains (+1.6 points) against the best transition-based AMR parser in very similar conditions. While using no graph re-categorization, our single model yields the second best Smatch score on AMR 2.0 (81.8), which is further improved to 83.4 with silver data and ensemble decoding.

Non-autoregressive (NAR) neural machine translation is usually done via knowledge distillation from an autoregressive (AR) model. Under this framework, we leverage large monolingual corpora to improve the NAR model’s performance, with the goal of transferring the AR model’s generalization ability while preventing overfitting. On top of a strong NAR baseline, our experimental results on the WMT14 En-De and WMT16 En-Ro news translation tasks confirm that monolingual data augmentation consistently improves the performance of the NAR model to approach the teacher AR model’s performance, yields comparable or better results than the best non-iterative NAR methods in the literature and helps reduce overfitting in the training process.

We propose an unsupervised method for sentence summarization using only language modeling. The approach employs two language models, one that is generic (i.e. pretrained), and the other that is specific to the target domain. We show that by using a product-of-experts criteria these are enough for maintaining continuous contextual matching while maintaining output fluency. Experiments on both abstractive and extractive sentence summarization data sets show promising results of our method without being exposed to any paired data.