arXiv subjects

Elliott Ash

Publications and source records attributed to Elliott Ash.

At least 19 recordsLinked to original sources

arXiv2026

Doppelganger: Sound Effects and Their Synthetic Twins

Audio-conditioned generators now produce synthetic sound effects from real recordings, so the real and synthetic versions of an event increasingly coexist in sound libraries and in the corpora used to train audio models -- yet no benchmark measures whether a representation can match a synthetic clip to the specific real recording it was generated from. I introduce Doppelganger, a benchmark for matching sound effects across the synthetic-real boundary, pairing 10,420 real clips across 34 everyday sound events each with an audio-conditioned synthetic twin, alongside a controlled 7-class corpus. Off-the-shelf audio encoders do not cross the boundary cleanly. Making the embedding ignore the boundary the standard way -- training it on sound-event labels -- works on familiar sounds but backfires on new ones, dropping below the untrained encoder. Training on the pairs instead -- a clip and its own synthetic twin -- generalizes. On sound events held out of training, it recovers the exact real source about 80% of the time (up from 61% untrained; chance 0.03%), whereas no objective meaningfully improves category-level recognition on those unseen events. The learned matching is specific to one generator -- it survives changes to that generator's settings but not a switch to a different generator, and collapses for the text-only generators tested. A human annotation baseline (49 listeners) lands well above chance but below the models on the same trials. Synthetic twins fool people into calling them real about 29% of the time, yet a generator-specific detector separates these audio-conditioned twins from real recordings perfectly.

cs.SD↗

arXiv2026

Characterizing Narrative Content in Web-scale LLM Pretraining Data

The narrative composition of web-scale LLM pretraining corpora remains largely unexplored, even though narrative is a fundamental mode of human communication. We present the first fine-grained study of narrative features in Dolma, a 3-trillion-token open pretraining corpus. Drawing on narrative theory, we design a framework spanning three core narrative elements (agency, setting, and events) operationalized as 11 interpretable dimensions. After curating and hand-annotating a diverse set of 400 passages, we create an LLM-labeled dataset of 25K passages, and finally, we finetune and validate NarraBERT, two RoBERTa-based models for fine-grained narrative prediction. We apply NarraBERT to 13M passages, resulting in a new dataset, NarraDolma. We find that narrative structure is measurable at scale across extremely heterogeneous data and narrative qualities are unequally distributed across pretraining sources, topics, and formats in ways that current data curation practices neither measure nor account for. Our framework, dataset, and analyses provide a foundation for understanding how narrative qualities are distributed in LLM pretraining data.

cs.CL↗

arXiv2026

Read the Paper, Write the Code: Agentic Reproduction of Social-Science Results

Recent work has used LLM agents to reproduce empirical social science results with access to both the data and code. We broaden this scope by asking: Can they reproduce results given only a paper's methods description and original data? We develop an agentic reproduction system that extracts structured methods descriptions from papers, runs reimplementations under strict information isolation -- agents never see the original code, results, or paper -- and enables deterministic, cell-level comparison of reproduced outputs to the original results. An error attribution step traces discrepancies through the system chain to identify root causes. Evaluating four agent scaffolds and four LLMs on 48 papers with human-verified reproducibility, we find that agents can largely recover published results, but performance varies substantially between models, scaffolds, and papers. Root cause analysis reveals that failures stem both from agent errors and from underspecification in the papers themselves.

cs.AI↗

arXiv2026

Good Question! The Effect of Positive Feedback on Contributions to Online Public Goods

Online platforms where volunteers answer each other's questions are important sources of knowledge, yet participation is declining. We ran a pre-registered experiment on Stack Overflow, one of the largest Q&A communities for software development (N = 22,856), randomly assigning newly posted questions to receive an anonymous upvote. Within four weeks, treated users were 6.3% more likely to ask another question and 12.9% more likely to answer someone else's question. A second upvote produced no additional effect. The effect on answering was larger, more persistent, and still significant at twelve weeks. Next, we examine how much of these effects are due to algorithmic amplification, since upvotes also raise a question's rank and visibility. Algorithmic amplification is not important for the effect on asking additional questions, but it matters a lot for the effect on answering other questions. The increase in visibility increases the probability that another user provides an answer, and that experience appears to shift the poster toward broader community participation.

cs.SI↗

arXiv2026

pdfQA: Diverse, Challenging, and Realistic Question Answering over PDFs

PDFs are the second-most used document type on the internet (after HTML). Yet, existing QA datasets commonly start from text sources or only address specific domains. In this paper, we present pdfQA, a multi-domain 2K human-annotated (real-pdfQA) and 2K synthetic dataset (syn-pdfQA) differentiating QA pairs in ten complexity dimensions (e.g., file type, source modality, source position, answer type). We apply and evaluate quality and difficulty filters on both datasets, obtaining valid and challenging QA pairs. We answer the questions with open-source LLMs, revealing existing challenges that correlate with our complexity dimensions. pdfQA presents a basis for end-to-end QA pipeline evaluation, testing diverse skill sets and local optimizations (e.g., in information retrieval or parsing).

cs.CL↗

arXiv2025

ReProbe: Efficient Test-Time Scaling of Multi-Step Reasoning by Probing Internal States of Large Language Models

LLMs can solve complex tasks by generating long, multi-step reasoning chains. Test-time scaling (TTS) can further improve performance by sampling multiple variants of intermediate reasoning steps, verifying their correctness, and selecting the best steps for continuation. However, existing verification approaches, such as Process Reward Models (PRMs), are computationally expensive and require large-scale human or model-generated annotations. We propose a lightweight alternative for step-level reasoning verification based on probing the internal states of LLMs. We train a transformer-based probe that uses the internal states of a frozen LLM to estimate the credibility of its reasoning steps during generation. Annotation can be provided either by a larger LLM (e.g., DeepSeek-R1) or in a self-supervised manner by the original model itself. The probes are lightweight, containing fewer than 10M parameters. Across multiple domains, including mathematics, planning, and general knowledge question answering, our probes match or exceed the performance of PRMs that are up to 810x larger. These results suggest that LLM internal states encode confidence in their reasoning processes and can serve as reliable signals for step verification, offering a promising path toward scalable, generalizable TTS and more introspective LLMs.

cs.AI↗

arXiv2025

Apertus: Democratizing Open and Compliant LLMs for Global Language Environments

We present Apertus, a fully open suite of large language models (LLMs) designed to address two systemic shortcomings in today's open model ecosystem: data compliance and multilingual representation. Unlike many prior models that release weights without reproducible data pipelines or regard for content-owner rights, Apertus models are pretrained exclusively on openly available data, retroactively respecting `robots.txt` exclusions and filtering for non-permissive, toxic, and personally identifiable content. To mitigate risks of memorization, we adopt the Goldfish objective during pretraining, strongly suppressing verbatim recall of data while retaining downstream task performance. The Apertus models also expand multilingual coverage, training on 15T tokens from over 1800 languages, with ~40% of pretraining data allocated to non-English content. Released at 8B and 70B scales, Apertus approaches state-of-the-art results among fully open models on multilingual benchmarks, rivalling or surpassing open-weight counterparts. Beyond model weights, we release all scientific artifacts from our development cycle with a permissive license, including data preparation scripts, checkpoints, evaluation suites, and training code, enabling transparent audit and extension.

cs.CL↗

arXiv2025

Measuring Scalar Constructs in Social Science with LLMs

Many constructs that characterize language, like its complexity or emotionality, have a naturally continuous semantic structure; a public speech is not just "simple" or "complex," but exists on a continuum between extremes. Although large language models (LLMs) are an attractive tool for measuring scalar constructs, their idiosyncratic treatment of numerical outputs raises questions of how to best apply them. We address these questions with a comprehensive evaluation of LLM-based approaches to scalar construct measurement in social science. Using multiple datasets sourced from the political science literature, we evaluate four approaches: unweighted direct pointwise scoring, aggregation of pairwise comparisons, token-probability-weighted pointwise scoring, and finetuning. Our study finds that pairwise comparisons made by LLMs produce better measurements than simply prompting the LLM to directly output the scores, which suffers from bunching around arbitrary numbers. However, taking the weighted mean over the token probability of scores further improves the measurements over the two previous approaches. Finally, finetuning smaller models with as few as 1,000 training pairs can match or exceed the performance of prompted LLMs.

cs.CL↗

arXiv2025

DialogueForge: LLM Simulation of Human-Chatbot Dialogue

Collecting human-chatbot dialogues typically demands substantial manual effort and is time-consuming, which limits and poses challenges for research on conversational AI. In this work, we propose DialogueForge - a framework for generating AI-simulated conversations in human-chatbot style. To initialize each generated conversation, DialogueForge uses seed prompts extracted from real human-chatbot interactions. We test a variety of LLMs to simulate the human chatbot user, ranging from state-of-the-art proprietary models to small-scale open-source LLMs, and generate multi-turn dialogues tailored to specific tasks. In addition, we explore fine-tuning techniques to enhance the ability of smaller models to produce indistinguishable human-like dialogues. We evaluate the quality of the simulated conversations and compare different models using the UniEval and GTEval evaluation protocols. Our experiments show that large proprietary models (e.g., GPT-4o) generally outperform others in generating more realistic dialogues, while smaller open-source models (e.g., Llama, Mistral) offer promising performance with greater customization. We demonstrate that the performance of smaller models can be significantly improved by employing supervised fine-tuning techniques. Nevertheless, maintaining coherent and natural long-form human-like dialogues remains a common challenge across all models.

cs.CL↗

arXiv2025

Co-DETECT: Collaborative Discovery of Edge Cases in Text Classification

We introduce Co-DETECT (Collaborative Discovery of Edge cases in TExt ClassificaTion), a novel mixed-initiative annotation framework that integrates human expertise with automatic annotation guided by large language models (LLMs). Co-DETECT starts with an initial, sketch-level codebook and dataset provided by a domain expert, then leverages the LLM to annotate the data and identify edge cases that are not well described by the initial codebook. Specifically, Co-DETECT flags challenging examples, induces high-level, generalizable descriptions of edge cases, and assists user in incorporating edge case handling rules to improve the codebook. This iterative process enables more effective handling of nuanced phenomena through compact, generalizable annotation rules. Extensive user study, qualitative and quantitative analyses prove the effectiveness of Co-DETECT.

cs.CL↗

arXiv2025

The Medium Is Not the Message: Deconfounding Document Embeddings via Linear Concept Erasure

Embedding-based similarity metrics between text sequences can be influenced not just by the content dimensions we most care about, but can also be biased by spurious attributes like the text's source or language. These document confounders cause problems for many applications, but especially those that need to pool texts from different corpora. This paper shows that a debiasing algorithm that removes information about observed confounders from the encoder representations substantially reduces these biases at a minimal computational cost. Document similarity and clustering metrics improve across every embedding variant and task we evaluate -- often dramatically. Interestingly, performance on out-of-distribution benchmarks is not impacted, indicating that the embeddings are not otherwise degraded.

cs.CL↗

arXiv2025

Can Reasoning Help Large Language Models Capture Human Annotator Disagreement?

Variation in human annotation (i.e., disagreements) is common in NLP, often reflecting important information like task subjectivity and sample ambiguity. Modeling this variation is important for applications that are sensitive to such information. Although RLVR-style reasoning (Reinforcement Learning with Verifiable Rewards) has improved Large Language Model (LLM) performance on many tasks, it remains unclear whether such reasoning enables LLMs to capture informative variation in human annotation. In this work, we evaluate the influence of different reasoning settings on LLM disagreement modeling. We systematically evaluate each reasoning setting across model sizes, distribution expression methods, and steering methods, resulting in 60 experimental setups across 3 tasks. Surprisingly, our results show that RLVR-style reasoning degrades performance in disagreement modeling, while naive Chain-of-Thought (CoT) reasoning improves the performance of RLHF LLMs (RL from human feedback). These findings underscore the potential risk of replacing human annotators with reasoning LLMs, especially when disagreements are important.

cs.CL↗

arXiv2025

LEXam: Benchmarking Legal Reasoning on 340 Law Exams

Long-form legal reasoning remains a key challenge for large language models (LLMs) in spite of recent advances in test-time scaling. To address this, we introduce LEXam, a novel benchmark derived from 340 law exams spanning 116 law school courses across a range of subjects and degree levels. The dataset comprises 7,537 law exam questions in English and German. It includes both long-form, open-ended questions and multiple-choice questions with varying numbers of options. Besides reference answers, the open questions are also accompanied by explicit guidance outlining the expected legal reasoning approach such as issue spotting, rule recall, or rule application. Our evaluation on both open-ended and multiple-choice questions present significant challenges for current LLMs; in particular, they notably struggle with open questions that require structured, multi-step legal reasoning. Moreover, our results underscore the effectiveness of the dataset in differentiating between models with varying capabilities. Deploying an ensemble LLM-as-a-Judge paradigm with rigorous human expert validation, we demonstrate how model-generated reasoning steps can be evaluated consistently and accurately, closely aligning with human expert assessments. Our evaluation setup provides a scalable method to assess legal reasoning quality beyond simple accuracy metrics. Project page: https://lexam-benchmark.github.io/.

cs.CL↗

arXiv2025

Balancing Truthfulness and Informativeness with Uncertainty-Aware Instruction Fine-Tuning

Instruction fine-tuning (IFT) can increase the informativeness of large language models (LLMs), but may reduce their truthfulness. This trade-off arises because IFT steers LLMs to generate responses containing long-tail knowledge that was not well covered during pre-training. As a result, models become more informative but less accurate when generalizing to unseen tasks. In this paper, we empirically demonstrate how unfamiliar knowledge in IFT datasets can negatively affect the truthfulness of LLMs, and we introduce two new IFT paradigms, $UNIT_{cut}$ and $UNIT_{ref}$, to address this issue. $UNIT_{cut}$ identifies and removes unfamiliar knowledge from IFT datasets to mitigate its impact on model truthfulness, whereas $UNIT_{ref}$ trains LLMs to recognize their uncertainty and explicitly indicate it at the end of their responses. Our experiments show that $UNIT_{cut}$ substantially improves LLM truthfulness, while $UNIT_{ref}$ maintains high informativeness and reduces hallucinations by distinguishing between confident and uncertain statements.

cs.CL↗

arXiv2024

A large language model-based approach to quantifying the effects of social determinants in liver transplant decisions

Patient life circumstances, including social determinants of health (SDOH), shape both health outcomes and care access, contributing to persistent disparities across gender, race, and socioeconomic status. Liver transplantation exemplifies these challenges, requiring complex eligibility and allocation decisions where SDOH directly influence patient evaluation. We developed an artificial intelligence (AI)-driven framework to analyze how broadly defined SDOH -- encompassing both traditional social determinants and transplantation-related psychosocial factors -- influence patient care trajectories. Using large language models, we extracted 23 SDOH factors related to patient eligibility for liver transplantation from psychosocial evaluation notes. These SDOH ``snapshots'' significantly improve prediction of patient progression through transplantation evaluation stages and help explain liver transplantation decisions including the recommendation based on psychosocial evaluation and the listing of a patient for a liver transplantation. Our analysis helps identify patterns of SDOH prevalence across demographics that help explain racial disparities in liver transplantation decisions. We highlight specific unmet patient needs, which, if addressed, could improve the equity and efficacy of transplant care. While developed for liver transplantation, this systematic approach to analyzing previously unstructured information about patient circumstances and clinical decision-making could inform understanding of care decisions and disparities across various medical domains.

cs.CY↗

arXiv2024

Variational Best-of-N Alignment

Best-of-N (BoN) is a popular and effective algorithm for aligning language models to human preferences. The algorithm works as follows: at inference time, N samples are drawn from the language model, and the sample with the highest reward, as judged by a reward model, is returned as the output. Despite its effectiveness, BoN is computationally expensive; it reduces sampling throughput by a factor of N. To make BoN more efficient at inference time, one strategy is to fine-tune the language model to mimic what BoN does during inference. To achieve this, we derive the distribution induced by the BoN algorithm. We then propose to fine-tune the language model to minimize backward KL divergence to the BoN distribution. Our approach is analogous to mean-field variational inference and, thus, we term it variational BoN (vBoN). To the extent this fine-tuning is successful and we end up with a good approximation, we have reduced the inference cost by a factor of N. Our experiments on controlled generation and summarization tasks show that BoN is the most effective alignment method, and our variational approximation to BoN achieves the closest performance to BoN and surpasses models fine-tuned using the standard KL-constrained RL objective. In the controlled generation task, vBoN appears more frequently on the Pareto frontier of reward and KL divergence compared to other alignment methods. In the summarization task, vBoN achieves high reward values across various sampling temperatures.

cs.CL↗

arXiv2024

Aligning Large Language Models with Diverse Political Viewpoints

Large language models such as ChatGPT exhibit striking political biases. If users query them about political information, they often take a normative stance. To overcome this, we align LLMs with diverse political viewpoints from 100,000 comments written by candidates running for national parliament in Switzerland. Models aligned with this data can generate more accurate political viewpoints from Swiss parties, compared to commercial models such as ChatGPT. We also propose a procedure to generate balanced overviews summarizing multiple viewpoints using such models. The replication package contains all code and data.

cs.CL↗

arXiv2024

DIRAS: Efficient LLM Annotation of Document Relevance in Retrieval Augmented Generation

Retrieval Augmented Generation (RAG) is widely employed to ground responses to queries on domain-specific documents. But do RAG implementations leave out important information when answering queries that need an integrated analysis of information (e.g., Tell me good news in the stock market today.)? To address these concerns, RAG developers need to annotate information retrieval (IR) data for their domain of interest, which is challenging because (1) domain-specific queries usually need nuanced definitions of relevance beyond shallow semantic relevance; and (2) human or GPT-4 annotation is costly and cannot cover all (query, document) pairs (i.e., annotation selection bias), thus harming the effectiveness in evaluating IR recall. To address these challenges, we propose DIRAS (Domain-specific Information Retrieval Annotation with Scalability), a manual-annotation-free schema that fine-tunes open-sourced LLMs to consider nuanced relevance definition and annotate (partial) relevance labels with calibrated relevance scores. Extensive evaluation shows that DIRAS enables smaller (8B) LLMs to achieve GPT-4-level performance on annotating and ranking unseen (query, document) pairs, and is helpful for real-world RAG development. All code, LLM generations, and human annotations can be found in \url{https://github.com/EdisonNi-hku/DIRAS}.

cs.IR↗