paperlist.ai

AI conference papers

AI conference paper library

Search papers from 30+ top conferences, including NeurIPS, ICLR, ICML, CVPR, ACL and AAAI. Filter by keyword, then open any paper for its abstract, authors and links.

arXiv

Subtract, Transport, or Replay? Auditable Deletion from Language-Model Memory

Vishwajith Ramesh

Exact deletion from persistent language-model memory depends on whether a record's effect remains addressable after later computation. Native Kimi Delta Attention (KDA) gives a negative result for the tested receipt interface: the corpus-pooled raw recurrent contribution changes by 12-49% with the suffix and remains 8-49% after a decay-ledger correction. Native omission also changes later transition and write terms and other active caches. Frozen-input transport succeeds on its fixed-input control; the changed terms place native omission outside the tested receipt classes. Checkpoint replay supplies the evaluated recomputation path; zero residual on final logits and all 80 audited KDA arrays verifies restoration across the declared checkpoint surface. The complementary result is constructive. We retrofit support-vector memory into frozen Gemma 3 without attention transfer, low-rank recovery, distillation, adapters, or language-model parameter updates. Prefix-mass preservation and one box per prefix solve give base-matched admission at 4B with 1.85% perplexity overhead. At 1B and 4B, verified deletion agrees with its conditional retained-key refit within 1.3e-10 maximum next-token KL; behavioral attacks at 4B reach never-stored or chance baselines. Across 1B, 4B, and 12B, the 4B checkpoint uniquely combines base-matched admission with low overhead. The paper's two contributions are a negative result for native KDA's tested receipt classes and a positive training-free construction for addressable pretrained memory.

arXiv

Similarity All The Way Up: Multilingual Generalization in LLMs Relies on Language-Level Similarity Structures

Supantho Rakshit, Adele Goldberg, Henry Conklin

As Large Language Models (LLMs) grow more capable across diverse tasks, their (in)ability to generalize remains difficult to quantify and poorly understood beyond limited domains. In particular, LLMs are known to struggle generalizing multilingually, to languages outside of English, and that are poorly attested in their training data. To understand why this may be, and what enables some models to perform better than others, we turn to a long history of work across the cognitive sciences, arguing that successful generalization derives from appropriate representations in similarity space. We look at how well LLMs' representations capture the hierarchical similarity structure between distinct languages. Strikingly, we show LLMs' latent representations largely recover the hierarchical structure of the Indo-European language family tree -- grouping languages that are members of the same subfamily closely together in representation space. Furthermore, we show that the degree to which models reflect the similarity structure of languages correlates with their performance on XNLI, a multilingual natural language inference benchmark. This extends classic work on similarity-driven generalization at scale, showing how models that represent similar languages similarly generalize better from one language to another.

arXiv

Evaluation design conditions the expert-vs-auto MeSH gap: a controlled comparison of bag-of-words and BiomedBERT on the Cohen benchmark

Samuel M. Okoe-Mensah

A systematic review begins with someone reading thousands of abstracts to identify the few that are relevant, and classifiers are used to prioritise that reading. Their inputs are often augmented with Medical Subject Headings (MeSH), assigned either by expert indexers weeks or months after publication or by automatic tools at once. We did not identify prior work comparing the two directly as classifier features, or asking whether that comparison's outcome depends on how the classifier is evaluated. Using the Cohen et al. (2006) drug-class benchmark, we compare expert assignment against one mechanical procedure, substring matching against a MeSH vocabulary drawn from the benchmark, across a bag-of-words logistic regression classifier (seven reruns) and BiomedBERT (five seeds) on three topics. Under the canonical 5-fold full-corpus design the bag-of-words gap on Statins is +0.096 WSS@95%. Stratified subsampling to matched corpus size (n=803) reduces it by roughly two thirds, to +0.033, with a bootstrap interval that includes zero; 10-fold cross-validation at full corpus size reduces it by roughly four fifths, to +0.021. BiomedBERT under canonical evaluation gives +0.020, a difference of 0.001 from the bag-of-words 10-fold result. An empirical power analysis on a single canonical run per topic indicates that a Statins-sized effect at the per-fold variances of the other two topics would not have been detectable at that design (MDE 0.254 for Opioids, 0.384 for ADHD); at the pooled fold count of the multi-run protocol the bound depends on an effective sample size the design does not determine. The results bound the specific lexical matcher tested rather than automatic MeSH indexing in general. More broadly, benchmark conclusions about feature sources can change substantially under reasonable changes to the evaluation design.

arXiv

Comparing Architectures for Supervised Political Scaling

Anna Golub, Sebastian Padó

Text scaling, the task of positioning political actors on an ideological scale, is a fundamental task in political analysis. To ease the need for manual analysis, various NLP methods have been proposed for this task, including classification- and regression-based approaches, showing successes as well as limitations. The goal of our paper is to consolidate the state of the art in this area. We ask two questions: (a) Can the performance of scaling methods be improved by predicting scales not individually but jointly? (b) Is there a middle ground between classification and regression?

arXiv

Topics as Proxies for Sociodemographics: How Conversational Context Affects LLM Answers

Vera Neplenbroek, Gabriele Sarti, Arianna Bisazza, Raquel Fernández

When large language models (LLMs) are used in high-stakes scenarios, such as legal, medical and financial advice, even a single conversation history is enough to drive differences in outcomes between users. Prior work has demonstrated that this results in outcome disparities between sociodemographic groups, with some groups receiving more advantageous outcomes than others. In this work, we demonstrate that LLMs actually struggle to infer user sociodemographics from a single conversation history and that although there are disparities between sociodemographic groups, they are minimal in magnitude. To investigate what is the main driver of disparities between users, we compare user sociodemographics to a range of (psycho)linguistic features of conversations, including conversation topic, emotions, and readability. We find that conversation topics are most predictive of LLM-generated advice within a conversational context, which, to some extent, function as proxies for sociodemographic groups and often affect advice in unpredictable ways. This is cause for concern and highlights the need for future research to better understand the effect of conversational context on LLM outputs in high-stakes scenarios.

arXiv

One prompt is not enough: Instruction Sensitivity Undermines Embedding Model Evaluation

Yevhen Kostiuk, Kenneth Enevoldsen

Instruction embedding models have become common among state-of-the-art models, however are evaluated using a single prompt per task. The single-point evaluation ignores a main problem of the instruction-based approach namely: sensitivity to the phrasing of the instruction. We present an empirical study of prompt sensitivity across 6 embedding models and 11 datasets. We show that reported scores misrepresent the distribution of scores over plausible prompts. The default prompt can both systematically understate or overstate performance. Furthermore, we show that the leaderboard ranking is not robust to prompt selection: a developer improve their rank by favorably selecting prompts, and under adversarial prompt selection any model can be promoted to first place. Our findings suggest that single-prompt evaluation is insufficient for instruction-tuned embedding models and that benchmarks should incorporate prompt robustness, either by evaluating over multiple prompts or by reporting sensitivity alongside point estimates.

arXiv

Robust Checkpoint Selection for Multimodal LLMs via Agentic Evaluation and Stability-Aware Ranking

Qinwu Xu, Zhuoheng Li, Jessie Salas

Selecting a final checkpoint for multimodal large language models (MLLMs) is challenging when late-stage candidates are closely matched and downstream evaluation signals are noisy. Small observed differences can be comparable to variability introduced by finite evaluation samples, LLM-based judges, and ambiguous multimodal evidence, while validation loss may not identify the checkpoint preferred by downstream evaluation. We formulate late-stage checkpoint selection as a stability-aware decision problem under evaluation uncertainty and propose a progressive framework combining pointwise filtering, listwise ranking, and pairwise refinement. Repeated evaluation-set subsampling is used to characterize ranking stability, while percentile-based aggregation accounts for lower- and upper-tail behavior. Experiments show that multimodal data evaluability is critical: quality-aware curation of OCR-heavy inputs reduces ranking flip rate from 32.5\% to 11.2\% and increases inter-run agreement from 0.61 to 0.84. We further observe divergence between validation-loss progression and downstream checkpoint preference in two independent MLLM settings. An additional public Qwen2.5-VL-7B reproduction across 11 checkpoints shows tightly clustered pointwise scores and frequently tie-dominated final pairwise comparisons, while repeated evaluation most often selects an intermediate rather than the final checkpoint. These results suggest that reliable MLLM checkpoint selection should quantify and reserve evaluation uncertainty rather than force decisions from small differences in a single metric.

arXiv

Natural Language Processing: A Comprehensive Practical Guide from Tokenisation to RLHF

Mullosharaf K. Arabov

This preprint presents a systematic, research-oriented practicum that guides the reader through the entire modern NLP pipeline --- from tokenisation and vectorisation to fine tuning of large language models, retrieval augmented generation, reinforcement learning from human feedback, prompt engineering, model compression, and multimodal systems. Seventeen hands-on sessions combine concise theory with detailed implementation plans, formalised evaluation metrics, and transparent assessment criteria. The work is not a conventional textbook: it is designed as a reproducible research artefact where every session requires publishing code, models, and reports in public repositories. All experiments are conducted on a single evolving corpus, and the work advocates open weight models over commercial APIs, with special attention to the Hugging Face ecosystem. The material is enriched by original research on low resource languages, incorporating linguistic resources for Tajik and Tatar --- subword tokenisers, embeddings, lexical databases, and transliteration benchmarks --- demonstrating how modern NLP can be adapted to data scarce environments. The practicum also covers advanced topics including AI agents, multi-agent systems, LLMOps, and efficient inference, preparing students for both research and industrial deployment. Designed for senior undergraduates, graduate students, and practising developers seeking to implement, compare, and deploy methods from classical ML to state of the art multimodal and agent-based systems.

arXiv

From Unstructured Recall to Schema-Grounded Memory: Reliable AI Memory via Iterative, Schema-Aware Extraction

Alex Petrov, Alexander Gusak, Denis Mukha, Dima Korolev

Persistent AI memory is often reduced to a retrieval problem: store prior interactions as text, embed them, and ask the model to recover relevant context later. This design is useful for thematic recall, but it is mismatched to the kinds of memory that agents need in production: exact facts, current state, updates and deletions, aggregation, relations, negative queries, and explicit unknowns. These operations require memory to behave less like search and more like a system of record. This paper argues that reliable external AI memory must be schema-grounded. Schemas define what must be remembered, what may be ignored, and which values must never be inferred. We present an iterative, schema-aware write path that decomposes memory ingestion into object detection, field detection, and field-value extraction, with validation gates, local retries, and stateful prompt control. The result shifts interpretation from the read path to the write path: reads become constrained queries over verified records rather than repeated inference over retrieved prose. We evaluate this design on structured extraction and end-to-end memory benchmarks. On the extraction benchmark, the judge-in-the-loop configuration reaches 90.42% object-level accuracy and 62.67% output accuracy, above all tested frontier structured-output baselines. On our end-to-end memory benchmark, xmemory reaches 97.10% F1, compared with 80.16%-87.24% across the third-party baselines. On the application-level task, xmemory reaches 95.2% accuracy, outperforming specialised memory systems, code-generated Markdown harnesses, and customer-facing frontier-model application harnesses. The results show that, for memory workloads requiring stable facts and stateful computation, architecture matters more than retrieval scale or model strength alone.

arXiv

Cat-DPO: Category-Adaptive Safety Alignment

Tiankai Yang, Yi Nian, Xinyuan Li, Ruiyao Xu, Henry Peng Zou, Kaize Ding, Xiyang Hu, Yan Liu, Yue Zhao

Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a single scalar that is applied uniformly to every preference pair. The result is a model that looks safe on average but stays relatively unsafe on a minority of harm categories. We cast safety alignment as a per-category constrained optimization problem and derive Cat-DPO, a direct-preference-optimization algorithm with a separate adaptive safety margin for each harm category. The margin tightens when the model still produces unsafe responses on a category and relaxes once the model catches up, so the training signal tracks each category's current difficulty rather than averaging under one global rate. Across two LLM backbones and six preference-learning baselines, Cat-DPO improves aggregate helpfulness and harmlessness and compresses per-category safety variance and the best-to-worst gap, offering a drop-in per-category refinement of direct preference safety alignment.

arXiv

Doctorina MedBench: A Dialogue-Based Benchmark and Evaluation Framework for Agent-Based Medical AI

Anna Kozlova, Stanislau Salavei, Pavel Satalkin, Hanna Plotnitskaya, Sergey Parfenyuk, Andy Nkansah

We present Doctorina MedBench, an evaluation framework for agent-based medical AI based on the simulation of physician-patient interactions. Unlike traditional medical benchmarks that rely on solving standardized test questions, the proposed approach models a multi-step clinical dialogue in which an AI system must collect medical history, analyze available synthetic attachments when present, formulate differential diagnoses, and provide diagnostic and management recommendations. System performance is evaluated across separate task-level domains, including diagnosis, differential diagnosis, treatment, safety-critical condition handling, and dialogue-step behavior; the broader D.O.T.S. framework is used as a supplementary summary for diagnosis, observations/investigations, treatment, and step count. The framework also supports testing and quality-monitoring workflows intended to identify changes in model behavior during development. It supports safety-oriented cases, category-based sampling of synthetic clinical scenarios, and regression-style comparisons across system versions. In the reported study, the analyzed paired complete-case cohort consisted of 254 physician-authored synthetic clinical cases retained from 261 attempted case identifiers. The evaluation metrics are intended for comparative research on interactive medical AI systems and for studying clinical reasoning workflows in synthetic dialogue settings. Our results suggest that simulated clinical dialogue can provide a complementary assessment setting to traditional examination-style benchmarks, while the reported findings do not establish independent clinical validity, clinical effectiveness, or readiness for real-world deployment.

arXiv

Masked diffusion LLMs can use EoS tokens for hidden reasoning

Sarah Breckner, Sebastian Schuster

Diffusion LLMs have been proposed as an alternative to autoregressive LLMs. Curiously, they are especially capable if the generation length, i.e., the number of tokens the model has to output, is set to a much higher value than the correct answer length, and the model pads its answer with end-of-sequence (EoS) tokens. We hypothesize that off-the-shelf masked diffusion LLMs use the representations of EoS tokens as additional computing capacity, which enhances their performance. We experiment with the diffusion models LLaDA1.5, LLaDA2.0-mini, and Dream-v0 on three reasoning tasks: Addition, Entity Tracking, and Sudoku. In a controlled prompting experiment, we confirm that adding EoS tokens improves the LLMs' performance. To further verify whether their representations are used for hidden computations, we perform a causal intervention and transfer the hidden states of the EoS tokens between generations, which increases the models' relative likelihood of outputting the counterfactual answer. The behavioral experiments and the causal interventions indicate that fully bidirectional masked diffusion LLMs can indeed perform latent reasoning in the representations of EoS tokens. Furthermore, we find that these results generalize beyond toy tasks and that providing the model with additional EoS tokens also improves performance on GSM8K and two-hop reasoning.

arXiv

QUIETT: Query-Independent Table Transformation for Robust Reasoning

Gaurav Najpande, Tampu Ravi Kumar, Manan Roy Choudhury, Neha Valeti, Yanjie Fu, Vivek Gupta

Real-world tables often contain schema inconsistencies, heterogeneous value formats, and implicit relational structures that degrade table reasoning and question answering. Existing approaches address these issues at query time, repeating normalization and restructuring for every new question and producing representations that do not generalize to unseen queries. We introduce QuIeTT, a transform-first, query-later framework that converts each raw table into a single SQL-ready canonical representation before any evaluation query is observed, under three constraints: no evaluation query access, information preservation, and single-table reuse. QuIeTT operates through a three-stage pipeline: issue probing, where synthetic queries surface structural deficiencies such as ambiguous schemas and heterogeneous formats; plan generation; and structured execution. Experiments on four benchmarks across five model families show consistent improvements over strong baselines in both direct prompting and agentic reasoning paradigms, with particularly strong gains on a manually annotated challenge set of structurally diverse unseen questions.

arXiv

Toward Federated Large Language Models in Medicine: A Parameter-Efficient Framework for Privacy-Preserving, Multi-Institutional Adaptation

Anran Li, Yuanyuan Chen, Wenjun Long, Yu Yin, Yan Hu, Hyunjae Kim, Weipeng Zhou, Yujia Zhou, Hongyi Peng, Yang Ren, Xuguang Ai, Zhenyue Qin, Ming Hu, Xiaoxiao Li, Han Yu, Yih-Chung Tham, Lucila Ohno-Machado, Hua Xu, Qingyu Chen

Large language models (LLMs) are increasingly adapted for medical applications, but most are trained using data from a single institution because privacy and governance constraints prevent multi-institutional data sharing. As a result, these models often generalize poorly across heterogeneous healthcare systems. We address this gap by introducing Fed-MedLoRA and Fed-MedLoRA+, a parameter-efficient federated framework for collaborative LLM adaptation across healthcare institutions. Fed-MedLoRA transmits only low-rank adapters rather than full model weights, reducing communication overhead. We also evaluate a privacy-preserving variant that applies Gaussian perturbation to transmitted adapter updates. Fed-MedLoRA+ further incorporates adaptive aggregation to better address cross-site heterogeneity in patient populations, annotation practices, and disease distributions. We evaluate the framework on clinical information extraction across five independent patient cohorts totaling 42,198 entities and 41,570 relations, and compare it with zero-shot and fine-tuned LLMs, domain-specific BERT models, and federated baselines. Across all settings, the proposed methods consistently improve extraction performance and generalize better to heterogeneous cohorts. In a real-world case study using clinical notes from the Yale New Haven Health System, the framework demonstrates strong performance under low-resource new-site deployment. These results suggest that federated, parameter-efficient LLM adaptation is feasible, scalable, and effective for multi-institutional clinical deployment.

arXiv

Learning to Diagnose and Correct Moral Errors: Beyond Shallow Heuristics in Moral Alignment

Bocheng Chen, Xi Chen, Han Zi, Haitao Mao, Zimo Qi, Xitong Zhang, Kristen Johnson, Guangliang Liu

Existing approaches to moral value alignment are primarily set out to align LLMs' generation with the distributions of morally appropriate language, which has seen good progress. However, these approaches are often brittle, heavily rely on shallow heuristics, and reduce performance in out-of-the-distribution tasks. In other words, the learning paradigm underlying existing approaches teaches LLMs what morally (in)appropriate language looks like, but not why it is morally (in)appropriate. In this paper, we address this challenge by developing pragmatic inference-driven methods to facilitate LLMs' learning of how to diagnose and correct moral errors, thereby enabling them to generate morally appropriate language. Pragmatic inference is the reasoning process of deriving (implied) meanings -- a famous concept in linguistics. Our methods vary the inference procedures by the inferential load of different moral discourses, rather than modelling their diverse and complex semantic distributions separately. Empirical results demonstrate that our approach improves moral value alignment in LLMs and generalizes effectively across tasks.

arXiv

CityRiSE: Reasoning Urban Socio-Economic Status in Large Vision-Language Models via Reinforcement Learning

Tianhui Liu, Hetian Pang, Xin Zhang, Jie Feng, Pan Hui, Yong Li

Urban socio-economic sensing plays a vital role in advancing global sustainable development goals. With the advent of Large Vision-Language Models (LVLMs), new opportunities have emerged to address this challenge by framing it as a multi-modal perception and reasoning task. However, recent studies show that LVLMs still struggle to make accurate and interpretable socio-economic predictions from visual data. To overcome these limitations and fully exploit the potential of LVLMs, we propose CityRiSE, a novel framework for Reasoning urban Socio-Economic status in LVLMs via reinforcement learning (RL). With carefully curated multi-modal dataset and verifiable reward design, our approach guides the LVLM to focus on semantically meaningful visual cues, enabling structured and goal-oriented reasoning for generalist socio-economic status prediction. Experiments demonstrate that CityRiSE, equipped with emergent reasoning, significantly outperforms existing baselines, improving both prediction accuracy and generalization across diverse urban contexts, especially on unseen cities and unseen indicators. This work highlights the promise of combining RL and LVLMs for interpretable and generalist urban socio-economic sensing.

arXiv

Gradual Code-Switching as Inference-Time Cross-Lingual Representational Alignment for LLMs

Haneul Yoo, Jiho Jin, Kyunghyun Cho, Alice Oh

While large language models (LLMs) have achieved notable progress in multilingual settings, their performance remains uneven across languages as LLMs often rely on English-centric latent representations. In this work, we introduce code-switching in-context learning (CSICL), an inference-time mechanism for cross-lingual representational alignment. Rather than relying on an abrupt translation pivot, CSICL explicitly scaffolds the reasoning trajectory by gradually transitioning from a target language to English, aligning non-English inputs with an English-centric reasoning space. Across 4 LLMs, 6 datasets, and 10 languages, CSICL consistently outperforms cross-lingual in-context learning baselines, yielding average gains of 6.0pp and 4.8pp in target and unseen languages, respectively. The improvements generalize across language families and are even more pronounced in low-resource settings, with gains of 14.7pp in target and 5.3pp in unseen languages. These findings establish code-switching as a robust and effective approach for reducing cross-lingual misalignment during inference, moving LLMs towards more equitable and effective multilingual systems.

arXiv

From Training to Generalization: Improving Moral Reasoning Through Pragmatic Inference

Guangliang Liu, Xi Chen, Bocheng Chen, Han Zi, Xitong Zhang, Kristen Johnson

Although moral reasoning has emerged as a promising research direction for large language models (LLMs), a persistent generalization challenge remains: LLMs often achieve strong performance on training data but struggle to generalize their moral reasoning to unseen test data. From a linguistic perspective, moral reasoning is a pragmatic process in which moral judgments are inferred based on the context of social norms underlying a given moral situation. However, existing approaches overlook this pragmatic nature because of two major bottlenecks: (1) LLMs are primarily skilled in capturing distributional semantics, which differs from the pragmatic nature of moral reasoning; (2) there is currently no effective solution for grounding language in the moral context. In this paper, we develop a pragmatic inference approach that enables LLMs to infer moral judgments for a given moral situation by combining metapragmatic links with Moral Foundations Theory. Specifically, metapragmatic links serve to bridge the gap between distributional semantics and pragmatics, whereas Moral Foundations Theory provides a principled basis for grounding language in moral contexts. Experimental results demonstrate that our approach substantially improves LLMs' generalization in moral reasoning, highlighting the potential of pragmatic inference for future moral reasoning research.

arXiv

SLAyiNG: A Diverse and Community-validated Dataset of Queer Slang

Leonor Veloso, Lea Hirlimann, Lucija Mihić Zidar, Philipp Wicke, Valentin Hofmann, Hinrich Schütze

Queer vernacular is rarely studied in NLP, despite advancements in resources and evaluation for other sociolects and informal language. Because of this, NLP systems often process queer language incorrectly, e.g., they misclassify it as hate speech or generate negative responses. To address this problem, we propose Slaying, the first real-world dataset of English queer slang. Slaying is community-validated, and includes over 500 queer slang terms that pertain to more than 20 queer subcommunities. We argue that queer language data resources have great potential in NLP -- e.g., as components of large pretraining corpora and as the basis for benchmarks -- and can improve queer users' experience of NLP systems. We leverage Slaying for two novel findings in support of this argument: (i) For a number of language models, we show that they are unbiased towards the queer community, but at the same time unable to process its language, i.e., absence of representation bias does not entail the absence of linguistic bias. (ii) Model performance on queer slang varies across queer subcommunities; it is generally worse for slang pertaining to African-American and Latine communities. These findings are relevant for both the queer NLP and the broader ML communities. Slaying is available to the public, and open to future revisions and extensions. Warning: This paper contains profane and potentially offensive language.

arXiv

REHEARSE: Experiential Rehearsal for Verbal Confidence Calibration in Large Language Models

Ke Fang, Tianyi Zhao, Qianwen Wang, Lu Cheng

Large language models (LLMs) often express verbal confidence that is poorly aligned with actual correctness, limiting their reliability in safety-critical applications. Existing prompt-based methods treat calibration largely as a one-shot inference problem, relying on either instance-level reasoning or post-hoc self-assessment. We introduce Rehearse (Experiential Rehearsal), a training-free method that instead enables models to adapt from their own scored confidence experience. In a credence-calibration game grounded in a strictly proper scoring rule, the model receives feedback on prior confidence decisions; this experience is summarized in a post-game trajectory prefix that captures systematic over- or under-confidence. At inference time, the model applies this cross-instance calibration signal to the chain-of-thought reasoning trace for each new question. Across four LLMs, three benchmarks, and five random seeds, Rehearse achieves the lowest average ECE among training-free methods with improved accuracy, reducing average ECE by 58% relative to the uncalibrated baseline. Code is available at https://anonymous.4open.science/r/Experiential-Rehearsal-7C77/.

Keep browsing

More papers from top AI conferences

30+ top conferences, by conference, year, and topic.