跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+1085新增论文
+884其中 AI 相关
+130已有论文修订
+94新增 AI 导读

arXiv 最近完成扫描:2026-11-19 06:59 · 全站最近入库:2026-11-19 06:59各信源进度 →

统计口径与数据新鲜度

统计 2026-11-18 07:37 — 2026-11-19 07:37(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-19 06:59最近发现新增版本:2026-11-19 06:59库内论文最新更新时间:2026-11-18 08:00

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

26,871 已收录论文22,375 AI 相关8,156 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 49 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-19 06:57(北京时间)
新论文 · v1arXiv:2611.22778

A Full-Ledger Counterfactual Audit of Adaptive Reasoning

Xiaopeng Chu, Wenzheng Kong, Wei Liu, Wenyi Zhang

原文摘要

Adaptive reasoning must decide whether to answer now or spend more compute. Predicting that the current answer is wrong does not establish whether further reasoning will improve it. We introduce the Full-Ledger Counterfactual Value Audit (FLCVA), which compares an immediate answer with four fixed-length continuations from the same frozen prefix under one answer procedure and accounts for the cost of every signal used by a policy. On GSM8K and MATH-500, a random 256-token continuation recovers a wrong answer in 26.6% of Qwen3-8B cases and 18.9% of DeepSeek-R1-Distill-Qwen-7B cases, while damaging a correct answer in 1.4% and 1.9%. Correctness labels from independently sampled continuations predict a separate continuation's outcome in both primary models. In offline replay, using these labels to allocate compute improves performance on DeepSeek; the Qwen3 gain remains inconclusive. Under full cost accounting, the tested lightweight, myopic controllers mainly recognize current errors and add little continuation-specific value beyond that shortcut. A Llama-8B replication and action/cost ablations examine the robustness of these findings. FLCVA separates measuring local continuation value, predicting it from observable signals, and using those predictions to allocate computation.

cs.AI · cs.CL

新论文 · v1arXiv:2611.21432

Just sort it! A note on the LP behind rank-one anti-loop edits

Matteo Fischetti

原文摘要

Weight-space programming removes the thinking loops of a small reasoning model by a rank-one edit W' = W + u v^T of an MLP output matrix, where u is a loop direction and v is the solution of a sparse Linear Program (LP) over the neuron activations. We investigate whether the LP is needed. We replace it by a sort: we take the 8 neurons whose activation has the smallest coefficient of variation, and give them equal shares of the push that the LP solution realizes. On Qwen3.5-0.8B the sorted edit scores 67.0% on 1000 problems (MATH-500 and GSM8K) against 67.4% for the LP edit and 9.5% for the unedited model; the paired difference is -0.4 points, with 95% confidence interval [-2.9, +2.1], inside a non-inferiority margin of 5 points declared before the runs. On three further sampling seeds the difference is -0.6 points, interval [-2.0, +0.8], and -0.8 points, interval [-2.3, +0.6], when the dose is set without the LP. The sort does not solve the LP, but the neurons it shares with the LP solution carry most of its mean push. The outcome is that, on this model, the edit acts as a nearly constant bias along u.

cs.LG · cs.CL · math.OC

新论文 · v1arXiv:2611.21303

Syllable- and Word-Level Evaluation of Tibetan OCR

Jacob Moore, Paula Lauren

原文摘要

Optical character recognition (OCR) for Tibetan is evaluated predominantly by character error rate (CER). Tibetan marks syllable boundaries but does not mark word boundaries, so a word error rate requires a segmenter. This study evaluates nine recognizer configurations on typeset, woodblock and two handwritten script styles, alone and in combination, and reports syllable- and word-level error rates, to our knowledge the first for Tibetan OCR. The configurations span CTC models, TrOCR encoder-decoders with and without post-training quantization, and a vision-language model, and each is scored under CER, the syllable error rate (SER) and three segmented word error rates (SWER). CER ranks systems largely as the higher-level rates do (Kendall $τ\geq 0.833$), with one pair ordered oppositely. On the effect of joint training, CER and SER record significant changes of opposite sign in 7 of 64 comparisons. SER needs no segmenter and ranks systems as SWER does under two of the three segmenters ($τ\geq 0.889$). The output of each configuration is also tagged for part of speech, annotated for named entities and machine translated, and each result is compared with that obtained on the reference text. As rough guidance, translation chrF of 60 is reached at SER of 0.047 or below and part-of-speech preservation of 0.8 at 0.078 or below, while no error rate separates the systems that reach entity preservation of 0.6 from those that do not.

cs.CL

新论文 · v1arXiv:2611.20126

A Corpus-Driven SentencePiece Unigram Tokenizer for Sindhi: Design and Benchmark Against Fourteen Existing Tokenizers

Ali Nawaz, Muqadas Fareed, Soonh Sadaf

原文摘要

Sindhi is a low-resource Indo-Aryan language written mainly in an extended Perso-Arabic script. Several Sindhi tokenizers have been released on public model hubs, but they have not been compared with one another on shared data. We train SentencePiece Unigram tokenizers with byte fallback at four vocabulary sizes (32k to 150k) on about 1.10 million Sindhi sentences and a smaller, repeated Urdu, Persian and Arabic slice, from four public sources, and compare them with fourteen existing tokenizers on four evaluation sets: FLORES+, Sindhi Wikipedia, a news-derived sentiment corpus, and a contemporary news corpus. The metrics are fertility, single-token retention, UNK rate, byte-fallback rate, and round-trip character loss. At 64k, our tokenizer averages 1.147 tokens per word, against 1.208 for the best existing tokenizer and 1.512 for XLM-RoBERTa, and keeps 88.6% of words as single tokens. On fertility, single-token retention and UNK rate it is best or tied for best in all four evaluation sets at 64k and above; at 32k it trails one baseline on FLORES+. Margins over the best Sindhi-adapted baselines are small (0.002 fertility on FLORES+ at 64k), and we report no significance tests. A round-trip check finds no character loss for our tokenizer and silent character deletion (up to 0.85% of characters) in two existing ones; seven others lose less than 0.1%. We evaluate intrinsic metrics only, so effects on downstream tasks remain untested. Code and tokenizer files are public (DOI: 10.5281/zenodo.23962474).

cs.CL

新论文 · v1arXiv:2611.19700

Where Should LLM Agents Adapt at Test Time? In Context, in Weights, or Both

Parsa Mirtaheri, Mohammadreza Armandpour, Keivan Alizadeh, Iman Mirzadeh, Parshin Shojaee, Aryo Lotfi, David Harrison, Mehrdad Farajtabar

原文摘要

LLM agents operating repeatedly in the same environment, such as a code repository, a database, or an application suite, encounter tasks that share latent structure. Experience can reveal this structure, allowing agents to reuse environment knowledge and learn from environmental feedback. Agents can adapt at test-time using this experience through two channels: retaining it in context and updating their weights. These channels are usually studied separately, leaving unclear what each contributes and when they complement each other. We compare in-context learning (ICL), reinforcement learning with verifiable rewards using GRPO, and self-distillation using SDPO, alone and in combination, using a shared online protocol and modular framework. Agents encounter matched task streams, are evaluated on each task before adapting to it, and learn only from their own attempts and environment feedback. We evaluate 14 adaptation mechanisms and a stateless baseline with Qwen3.5-9B on 109 tasks from four environments drawn from SWE-bench Verified, AppWorld, ScienceWorld, and BIRD-Critic-SQLite, using five task orders per environment. ICL provides the largest standalone gain: its best policy raises mean success across environments from 28.8% to 49.4%. Adding GRPO to this policy achieves the highest mean success, 51.6%. While adding GRPO improves the overall mean under every ICL policy, SDPO has mixed effects. Trace analyses show fewer lookups and shorter attempts under ICL, consistent with reusing knowledge of the shared structure. However, ICL also frequently skips inspection and verification for the current task. Adding GRPO to ICL helps restore these behaviors while preserving its reduced environment lookups. These observations suggest that reinforcement learning with verifiable rewards can refine how agents use retained experience, motivating the joint study of context management and weight updates.

cs.LG · cs.AI · cs.CL

新论文 · v1arXiv:2611.17939

Can You Trust Your Finance Agent? Drift, Diagnosis and Auditable Intervention on One Shipped Financial Research Agent

Cosimo Galeone, Anna Ettorre, Ginevra Mariani, Minsu Park, Giuseppe Ettorre, Daniele Ligorio

原文摘要

Can you trust a finance agent you did not build? We take the reference agent shipped with the Vals Finance Agent Benchmark v2, leave it untuned, and ask four questions a leaderboard score does not answer. Is the score stable? Seven identical runs on 18 questions pass between 105 and 123 of 149 rubric criteria, and the bill varies by up to 17%. When the agent fails, who owns the fix? We classify 20 failed trajectories, from three runs of nine held-out questions, by the layer whose owner would remedy them, under a taxonomy frozen in advance: 13 belong to the agent's instructions, 3 to its tools, 4 to the model, none to its data. Can that diagnosis be trusted? Four blind readers from three model families, given only the traces, name the same fault as our LLM analyzer in 94% of cases, the rate at which they agree with one another (93%). Can a change be trusted? An optimization loop reads the traces, makes one targeted change per round and keeps it only if it clears a fixed threshold. Under a cost objective its final configuration costs 0.34x the baseline on the held-out questions (95% CI [0.22, 0.58]) at accuracy within noise. Under an accuracy objective it gains 14 points on the development questions and 1.5 on the held-out ones (CI [-5.3, +7.2]). Judged against the measured run-to-run noise, under a rule written before the test was opened, the first is a real saving and the second is not. This is a case study of one agent and nine test questions, not a general result.

cs.CL · cs.AI

版本更新 · v4arXiv:2610.23170

Chinese Decision: A Family of Language Models for Single-Pass Probabilistic Decision-Making

Qiang Yan, Jiafeng Guo

原文摘要

Many applications need either a probability distribution over a few decisions or a set of typed spans from a text, and both are usually obtained by generating text. We present Chinese Decision, a family of Chinese models that answers both with a single forward pass and no generated token: candidate-letter logits give probabilities for Boolean, categorical, and ordinal decisions, and a small span head reading an echoed copy of the input gives typed character spans. Both readouts share one backbone and one merged LoRA, so a unified model is one weight file. We release decision-only and unified models at 0.8B, 2B, and 4B (Qwen3.5), trained for one epoch on 13.05 million decisions and, for the unified models, on CJ-IE, a 1.2-million-row Chinese extraction mixture. On the full 307,900-decision CJ-Bench, the unified 4B model reaches 86.86% accuracy, above every reported baseline and the decision-only 4B model (85.97%), and extracts spans on CJ-IE with micro F1 of 68.9. Retention has a price: against a 0.8B model continued on CJ-IE alone, which reaches 73.9 F1 but forgets CJ-Bench (71.02%), keeping decision ability costs 5.0 to 10.7 F1 points, shrinking with size. Gains of the unified models over decision-only ones are small, and all comparisons rest on single training runs with paired clustered bootstrap intervals.

cs.CL

新论文 · v1arXiv:2611.15049

Science Prediction Universe

Zengji Tu, Xiang Liu, Jiarui Sun, Ruoyu Wu, Zhihui Qi, Chengda Wu, Yutao Zhou, Bo Yang 等 36 位作者

原文摘要

Humanity has long sought to explain natural phenomena. Reliable prediction of experimental outcomes is a central test of our understanding of the processes that produce them. If artificial general intelligence (AGI) is to surpass human scientific capabilities, it must also predict experimental outcomes more reliably than humans. Such capabilities could accelerate scientific discovery across disciplines, help address humanity's most pressing challenges, and improve human well-being. We introduce Science Prediction Universe (SPU), a benchmark of hidden experimental outcomes with 100 public verified tasks from chemistry, materials science, biology and medicine, and 16 private tasks from unpublished wet-lab experiments. Given an experimental setup and candidate conditions, an agent works with any resources it needs to rank the candidates by a measured outcome. We evaluate thirteen agent configurations spanning twelve frontier models across two harnesses (Codex and Claude Code). Current agents remain far from saturating wet-lab outcome prediction, with substantial performance disparities across models: on the public benchmark, configurations separate into two distinct tiers, with task-adjusted Spearman correlations ranging from 0.227 for the lowest-performing configuration to 0.478 for the best. Analysis of 1,155 scored public trajectories (227,103 tool calls) suggests that the bottleneck lies in scientific judgment rather than tool execution: agents rely on shared monotone priors, agreeing far more with one another than with experimental reality (0.52 vs. 0.33), and frequently compute trivial physical or chemical proxies that do not influence the final outcome.

cs.AI · cs.CL

新论文 · v1arXiv:2611.14935

Do Large Language Models Show Racial and Sex Bias in Pain Management? A Counterfactual Audit of GPT, Claude, Gemini, Llama, DeepSeek and Qwen

Richard J. Young, Alice M. Matthews

原文摘要

Racial and ethnic minority patients receive less pain treatment than non-Hispanic White patients, and large language models (LLMs) used as medical AI could reproduce that bias. Prior audits tested a few proprietary models, stated race explicitly and rarely measured the models' own response instability. In exploratory, non-preregistered analyses, Study 1 had four open-weight LLMs assess six acute-pain vignettes in every combination of four races and two sexes, with identity stated or cued only by a name (two names per group), comparing each identity with a White male reference on the same case and sampling seed. In Study 2, 24 configurations of 17 models, including OpenAI's GPT-4.1 mini and gpt-oss, Anthropic's Claude Haiku 4.5, Google's Gemini 2.5 Flash, Meta's Llama, DeepSeek and Qwen, with reasoning on and off and at two precisions, answered the same vignettes and the 50 clinician-written Q-Pain cases. In this single-prompt, JSON-constrained task, no difference survived correction in Study 1, none of 1,386 comparisons survived the false discovery rate in Study 2, and every pain-severity difference was equivalent to zero within one point on a 0-10 scale (within a quarter point for 237 of 238 Q-Pain comparisons). Opioid decisions were bounded less precisely (Study 1 could not exclude differences of about 20 percentage points), so they show no detected difference, not proof of none. The few differences surviving correction within a model favoured Black patients. Models differed far more from one another and across their own settings than between patients (the reference patient received the high opioid dose in 6% to 93% of responses), although serving software, provider and prompt format also differed. Fairness audits of clinical AI should use several names per group, pair comparisons on the seed, report the noise floor and test each deployed configuration.

cs.CL · cs.AI · cs.CY

新论文 · v1arXiv:2611.13843

PolyRoute: A Twelve-Language and Code-Mixed Benchmark for Typed-Decision Routing, with Baselines and a Leaderboard

Mohammed Nayeem Khan, Sarun Kumar K S

原文摘要

Typed-decision encoders answer a question about a text in one forward pass, but their behaviour outside English is rarely measured on a fixed yardstick. We release PolyRoute: 322 native-language support messages in 12 languages and 9 writing systems, 156 code-mixed or romanised messages (Hinglish, Spanglish, Arabizi), 967 optional native training rows, and a protocol for a three-way department-routing decision. Labels come from Qwen3.8-27B with a blind second pass (91% of candidates kept); an independent model, gemma4-31b, agrees on 97.2% of test items, and the largest character 5-gram similarity between any test text and any training text is 0.33. A leaderboard of 13 systems gives 61.2% for the English-root checkpoint, 77.3% for the shipped multilingual one and 96.6% for the best fine-tune; TF-IDF with logistic regression reaches 91.9%. On four withheld languages the three multilingual systems score 77%, 90% and 93%. On code-mixed text the English root (85%) is nominally ahead of the shipped multilingual checkpoint (79%; p=0.134). The split resolves differences only above about 4 points. All texts and labels are LLM-generated.

cs.CL · cs.AI

新论文 · v1arXiv:2611.13216

Latin American Studies since the 1980s: From Apocalyptic Dreams to Assimilationist Consolidation

Thiago Dumont Oliveira, Juan Grigera, Pedro Mendes Loureiro

原文摘要

We investigate how Latin American Studies (LAS) has evolved since 1982 in three leading LAS journals. Using network analysis, text mining, and bibliometrics, we trace transformations in disciplinary positioning, authorship, and citation patterns. Our main claim is that LAS has undergone an assimilationist consolidation. The field is now institutionally recognisable, with core LAS journals occupying central positions in disciplinary networks and acting as bridges across fields. Yet this consolidation has reinforced hierarchies privileging Anglophone academic norms. Journals based in Latin America, once central, have been marginalised. The share of Latin American authors has grown, but their profiles converge towards Northern standards.

cs.DL · cs.CL

新论文 · v1arXiv:2611.13207

HiCAP: Hierarchical Cultural Adaptation via Personality for Dialogue Generation

Lingzhi Shen, Xiaohao Cai, Yunfei Long, Imran Razzak, Shoaib Jameel

原文摘要

Cultural adaptation in dialogue systems refers to the ability to generate responses that conform to linguistic norms and social expectations across different cultures. However, existing dialogue generation approaches pay limited attention to the role of personality in shaping cultural expression and communication style. We therefore propose HiCAP, a novel framework that integrates personality traits into culturally adaptive dialogue generation. HiCAP employs a personality-conditioned control mechanism that hierarchically modulates response style by first aligning it with the cultural tone and then refining linguistic expression and content according to individual personality attributes. To embed these factors, the model combines supervised fine-tuning and instruction-based tuning, supported by dynamic instruction templates and multi-agent demonstrations to capture diverse combinations of culture and personality. Experiments on CulturalPersonas and CPED show that HiCAP achieves the strongest overall performance across various dialogue generation metrics, demonstrating effective cultural adaptability and personality modelling.

cs.CL · cs.LG

版本更新 · v6arXiv:2507.06565

A Mathematical Theory of Discursive Networks

Juan B. Gutiérrez

原文摘要

Large language models (LLMs) turn writing into a live exchange between humans and software. We characterize this new medium as a discursive network that treats people and LLMs as equal nodes and tracks how their statements circulate. We define the generation of erroneous information as invalidation (any factual, logical, or structural breach) and show it follows four hazards: drift from truth, self-repair, fresh fabrication, and external detection. We develop a general mathematical model of discursive networks that shows that a network governed only by drift and self-repair stabilizes at a modest error rate. Giving each false claim even a small chance of peer review shifts the system to a truth-dominant state. We operationalize peer review with the open-source \emph{Flaws-of-Others (FOO) algorithm}: a configurable loop in which any set of agents critique one another while a harmonizer merges their verdicts. We identify an ethical transgression, epithesis, that occurs when humans fail to engage in the discursive network. The takeaway is practical and cultural: reliability in this new medium comes not from perfecting single models but from connecting imperfect ones into networks that enforce mutual accountability.

cs.CL · cs.LG

新论文 · v1arXiv:2611.11316

ReGKG-2025: A GKG-Compatible Reconstruction of the GDELT Global Knowledge Graph for the Missing Period of 14 June to 2 July 2025

Nikita A. Klimin

原文摘要

The June 2025 outage in the Global Database of Events, Language, and Tone (GDELT) collection pipeline left a gap in the Global Knowledge Graph (GKG), interrupting article-level news metadata used in downstream research. ReGKG-2025 is an archive-derived sample of 281,620 reconstructed article records covering all 1,665 missing 15-minute batches. Articles were recovered from Common Crawl and the Wayback Machine using source quotas estimated from surrounding native GKG data. Text was screened and normalized, non-English bodies were translated, and flagged records were manually reviewed. Eligible articles were converted to the 27-column GKG 2.1 layout. The versioned package includes record-level provenance, archive replay and WARC locators, source summaries, source-by-day calibration weights, and the complete manual decision journal. The sample reaches about 14% of modeled native volume and represents sources accounting for 84% of selected reference record mass. In an intact-week test, 41% of recovered articles matched native GKG records; field agreement on matched pairs is close to the processor baseline (tone MAE 1.69). The dataset supports retrospective content analysis and source-level aggregates; its tone and theme distributions differ from native GKG, and it does not support exhaustive counts. Data: https://doi.org/10.5281/zenodo.23964265.

cs.DL · cs.CL

版本更新 · v2arXiv:2610.36617

Balanced State Subspaces with Spectral Rank Allocation for Mamba-2 State Compression

Shang Tingyu, Cao Wenfei, Han Zhi

原文摘要

Mamba-2 compresses all context into a fixed-size recurrent state that dominates decoding memory and bandwidth. Existing compressors quantize state values or prune coordinates with per-channel scores. Trained states are nearly rank-one and bias-like, aligned with the mean write; but the output's energy spans an order of magnitude more directions. Per-channel scores see only the diagonal of the write--read operator, so they cannot select the subspace that governs the input--output map, which is a rotation of the coordinates. We therefore propose \method{}, which keeps that subspace by applying balanced truncation exactly via Mamba-2's scalar-identity transition. Layers differ just as sharply: most are bias-dominated, while the few that carry long-range associations are the flattest and need most of their state. We therefore introduce \spar{}, which redistributes a fixed total rank budget across layers according to the Hankel spectrum. On Mamba-2 checkpoints from 130M to 2.7B, \method{}+\spar{} halves the state at an average cost of $0.3$ WikiText-2 perplexity points, against $1.6$ for GHOST, while retaining $97\%$ of the dense model's five-digit in-context recall. At a quarter of the state, giving every layer the same rank keeps $5\%$ of that recall and \spar{} $77\%$, for a $7\%$ perplexity cost.

cs.LG · cs.AI · cs.CL · stat.AP · stat.ME

版本更新 · v3arXiv:2607.25959

Detecting Knowledge Inconsistencies Across Text, Tables, and Knowledge Graphs

Fanfu Wei, Thibault Ehrhart, Raphaël Troncy

原文摘要

Wikipedia and Wikidata are widely used for information access, LLM pre-training, and retrieval-augmented generation. Their knowledge is deeply connected but scattered across text, tables, and knowledge graphs. This raises a practical question: when these modalities disagree, how can we detect and explain the conflict? We study this problem as modality-level inconsistency detection. We first introduce a taxonomy of cross-modal knowledge inconsistencies, covering information granularity differences, direct conflicts, temporal changes, and KG incompleteness. We then present Kontrast, an automatic framework that uses Text-to-SPARQL and LLM reasoning to compare table-based answers with KG evidence and categorize the resulting inconsistencies. Experiments on various Table-QA datasets show that cross-modal inconsistencies are common and informative. They reveal not only true knowledge conflicts, but also missing KG structure and temporal mismatches while being limited by Text-to-SPARQL errors and noise. Our analysis shows that text, tables, and KGs can complement and correct one another through systematic comparison. Kontrast provides a practical tool for large-scale knowledge auditing and establishes a benchmark for future work on cross-modal knowledge consistency. Code and data are available at https://github.com/ECLADATTA/KONTRAST.

cs.CL · cs.AI

版本更新 · v2arXiv:2611.05730

The Male Abstraction Bias in Internet Culture and Large Language Models

Douglas Guilbeault, Bhargav Srinivasa Desikan

原文摘要

Conceptual abstraction and gender are two equally central aspects of the human mind, yet they have traditionally been studied in isolation from one another, as though they were fundamentally separate. Here, we find gender and abstraction are systematically linked throughout the space of concepts spanning single (N = 40,000) and multi-word expressions (N = 67,000): across billions of words and millions of images on the internet --- and across numerous measurement techniques, including human and AI classifications --- more abstract concepts are more associated with men, a pattern we show replicates in languages around the world. This male abstraction bias not only holds across semantic domains, but it also maps onto grammatical structure: verbs and adverbs are more abstract and more male-coded, revealing a system-wide trace of stereotypes accentuating agentic qualities in men. Furthermore, we show this bias readily emerges in the free associations humans and AIs produce in controlled online tasks, indicating its psychological internalization and its algorithmic reproduction. These findings identify a latent structure in conceptual space that unifies a range of stereotypes: for example, framings of women as more embodied, emotional, and warm, versus men as more cognitive, rational, and competent; as well as erroneous views of men as better at mathematics, science, and engineering. More broadly, by analyzing an unprecedentedly large, multimodal sample of online data, we provide culture-wide evidence of gender as a core organizing principle shaping conceptual space far beyond the characterization of people and social roles.

cs.CY · cs.CL · cs.HC

版本更新 · v2arXiv:2610.37019

V-LUMEN: Visual Lookup Memory for Embedding Scaling in Vision-Language Models

Miso Choi, Daekeun Kim, Eunji Kim, Jungbeom Lee

原文摘要

Embedding-level scaling expands the capacity of language models through lookup-based memory rather than by increasing model depth, width, or dense computation. Extending this idea to vision-language models (VLMs) is challenging because visual features are continuous, spatially structured, and query-dependent, unlike discrete text tokens that naturally support n-gram lookup. We introduce V-LUMEN, a visual lookup memory for VLMs. V-LUMEN discretizes visual features with a frozen codebook, builds spatially aware memory keys through spatial aggregation-based hashing, and modulates them with text-conditioned relevance signals; all addressing is non-parametric, and the retrieved embeddings are fused into the backbone through a gated residual pathway. On five reasoning-intensive benchmarks, V-LUMEN substantially improves a frozen Qwen3-VL-2B backbone and transfers zero-shot to three benchmarks whose images are disjoint from the training data. At a matched parameter budget, moving part of a LoRA or MoE adapter's capacity into V-LUMEN preserves accuracy on two backbones, indicating that external visual memory is an independent capacity axis rather than a stronger adaptation method. Because lookup cost is constant in table size and incurred only once at prefill, this capacity grows in storage rather than in per-token compute. Code is available at https://github.com/miso-choi/V-Lumen.

cs.LG · cs.AI · cs.CL · cs.CV

新论文 · v1arXiv:2611.06655

Beyond accuracy: what AI assurance needs to establish in public services

Louise Yorke

原文摘要

Public bodies increasingly rely on AI-produced evidence, including classifications, summaries and assessments. These outputs can inform policy decisions and service delivery, making their reliability a matter of public accountability. Establishing whether this evidence can be relied upon requires testing its outputs and examining how organisations decide whether those tests are sufficient. We combine technical evaluation with analysis of AI governance requirements, informed by measurement theory and management-based regulation. We compare eight GPT-4o configurations classifying 1,200 UK Prime Minister's Questions into 21 policy topics, against human-coded labels and across unchanged repeats. Accuracy ranges from 68.2% to 70.3%, yet 24.3% of questions receive different labels, with changes in topic shares and rankings. Configurations agree more with one another than with human coding: agreement does not guarantee correctness. Under the implemented score-extraction procedure, the proportion flagged for human review ranges from 0.9% to 41.0%. We analyse 37 UK and EU policy documents. Both sets expect testing and oversight, but reporting expectations do not consistently cover changing answers, shared errors, selective evaluation, substantive conclusions or review routing. Some UK guidance explicitly addresses stability. The UK sample emphasises public bodies checking their own use; the EU places substantial testing duties on providers while leaving them important choices about measures and acceptance thresholds. Without appropriate checks, evidence could misrepresent policy issues, redirect attention or resources, and leave errors undetected. Public policy should require organisations to justify reliance through checks beyond aggregate accuracy, clear responsibility, access to expertise and scrutiny of results.

cs.CY · cs.CL

版本更新 · v3arXiv:2610.33004

Entity-Aware Evaluation of OCR Systems for Historical Document Transcription

Camilo Mariño, Diego Belzarena, Marina Gardella, Ignacio Ramírez, Gregory Randall, Jean-Michel Morel

原文摘要

Optical Character Recognition (OCR) is essential for digitizing historical archives, yet standard evaluation metrics such as Character Error Rate (CER) do not necessarily reflect whether meaningful information, such as names, dates, and identifiers, is preserved. In this work, we investigate entitylevel analysis of OCR errors as a complementary perspective for evaluating OCR quality in historical documents. Using a publicly available benchmark of 175 documents from the Berrutti Archive, augmented with 5,209 manually annotated entity mentions that will be released upon acceptance, we compare sequence-recognition OCR systems, OCR-specialized vision-language models, and general-purpose vision-language models. Beyond global transcription accuracy, we analyze entity preservation and recoverability under varying character-level edit tolerances. Our experiments show that OCR rankings depend on the evaluation criterion: while Qwen 3.6 achieves the highest entity coverage across all tested tolerances, several systems with lower exact-match coverage recover substantially when small errors are tolerated, suggesting that many of their errors correspond to localized character-level deviations rather than fundamentally incorrect entity transcriptions. Conversely, some systems remain unable to recover many entities even under relaxed matching. These findings show that OCR quality cannot be fully characterized by aggregate transcription metrics alone and highlight the importance of evaluating how OCR errors affect the preservation and recoverability of meaningful information.

cs.CL · cs.AI · cs.CV · cs.LG · eess.IV

新论文 · v1arXiv:2611.05340

MedPhysBench:评估语言模型在医学物理计算与错误检测上的基准

MedPhysBench: A Benchmark for Evaluating Language Models on Medical Physics Calculations and Error Detection

Omar Nusrat

AI 速读 基于摘要

作者构建了一个覆盖放疗、诊断影像、核医学与辐射防护的医学物理计算与查错基准,并报告了12个语言模型在留出测试集上的表现。

physics.med-ph · cs.AI · cs.CL

查看原文摘要

Medical physicists turn a prescription into a delivered dose and check one another's calculations, because one error in a beam calibration or a monitor-unit (MU) calculation can affect every patient treated with it. MedPhysBench has ten seeded generators of calculation items in radiation therapy, diagnostic imaging, nuclear medicine and radiation protection, up to clinical tasks such as a TG-51 calibration worksheet. Reference answers come from code implementing the governing protocols, and a grader converts units and applies tolerance bands. From the same inputs, separate checking items test error detection, showing a colleague's worksheet or only the colleague's result, half correct and half with one planted error. On a held-out test split of 1,200 items (half for two models), 12 models scored 33.3 to 100.0% on calculation. Of the nine answers graded incorrect for the four best (99.3-100.0%), five contained model errors and four followed NCRP Report 151 on points our items misstate or leave open. Below them, most models did worst on level-3 clinical tasks, and Claude Sonnet 5.5 omitted the inverse-square correction in 19 of 20 isocentric MU items. Five models detected at least 96.7% of mostly large planted errors with at most 4% false alarms; for four of the other seven, checking only the result raised false alarms significantly. In an exploratory targeted test with 20 public cases of an omitted inverse-square correction, the three best models tested caught it in every worksheet, and three weaker models accepted 16 to 18 of 20 planted worksheets, one of them while applying the correction in most of its own calculations. Claude Opus 5.5 accepted 6 of 20 wrong results, although it applied the correction in all 20 of its own. We report the effect of two problems in our shielding items, found after the runs, and release the code, per-item outcomes and targeted-test responses.

新论文 · v1arXiv:2611.05292

人格设定不是人:身份是状态而非属性

Personas Are Not People: Identity as State, Not Attribute

Tamunotonye Harry, Yuanyuan Feng

AI 速读 基于摘要

作者论证基于固定人格设定的合成人群无法复现人群在具体情境中的行为,并提出以情境中的人为核心的新评估标准与设计。

cs.CL · cs.AI · cs.CY

查看原文摘要

Synthetic populations, large sets of simulated people, are increasingly used in place of real people to pretest surveys, simulate users and communities, and evaluate AI systems. Each simulated person is defined by a persona, a written profile of fixed traits, demographics, and attitudes that a language model is instructed to act out. Every such system assumes that a person's identity is a stable attribute that behaves the same way in every situation the simulation presents. We show that this assumption is not a harmless simplification. Three decades of personality research find the opposite. How a given person behaves varies more across contexts than people vary from one another. Drawing on identity economics, which describes how different situations activate different identities and norms within the same person, we separate behavior into three parts, which are stable traits, the identity a context calls up, and the situation itself. We then prove, under two stated assumptions, that any simulation built on fixed personas, no matter how detailed the personas or how large the underlying model, cannot reproduce the part of a population's behavior that is specific to it and its circumstances. We propose a new evaluation criterion based on this result and a person-in-context design that can meet it in principle. We test one version on a naturalistic Reddit personality corpus and on a published clinical-trial recruitment survey of a community with documented distrust of medical institutions. The design does produce context-dependent behavior when tested directly, but in neither setting does it fully match the real population, and we show where it falls short. The consequence for practice is a risk that synthetic-population evaluations misrepresent the groups whose behavior is most shaped by context, a distortion that richer personas cannot correct.

版本更新 · v2arXiv:2610.24715

When Do Repetition Loops Lock? Truncation Sampling, Bounded Penalties, and a Blind Spot in DRY

Yichuan Chen, Xin Wang

原文摘要

Language models sometimes repeat verbatim until the generation limit. We study the exit rather than the entrance: given that a span has already been repeated $n$ times, when is escape still possible? Along the loop the context is deterministic, so the probability that a loop survives $m$ more repetitions is, under any sampler, a product of per-step continuation probabilities computable from one forward pass. A lock, where this probability stays exactly one, is its limiting case. Pure sampling never locks; top-$p$ with threshold $π>1/2$ locks a position exactly when the continuation probability reaches $π$. A bounded penalty can leave a position locked; a penalty that outgrows the model's preference for the loop over unseen tokens guarantees escape under top-$p$. The same analysis shows a blind spot in the widely deployed DRY sampler: its sequence breakers cap the penalty at $λb^{g-1-a}$ for loops whose longest breaker-free run has $g$ tokens. On eleven open models, 61% of locked units with $g=6$ stay locked under DRY and 0.2% with $g=8$. In free generation with llama.cpp, all 199 loops that survive DRY, on six models, have $g\le 7$. We propose a count term that grows exponentially with the number of verbatim recurrences, up to its look-back window, and closes this blind spot. Added to DRY (DRY-C), it cuts loops in 1,800 structured generations on five models from 167 to 34, and loops reaching the length limit from 150 to 0, with reasoning accuracy and item coverage on par with DRY. DRY's own length term costs 2.5 to 14.2 points of accuracy on GSM8K and MATH-500; used in its place, the count term keeps the default sampler's accuracy and leaves one of the 1,800 generations at the limit.

cs.CL · cs.LG

版本更新 · v2arXiv:2610.26730

EpiLens: Lightweight RAG for Epidemiological Knowledge Acquisition

Vincenzo Perri, Ciro Cattuto, Daniela Paolotti

原文摘要

To prepare for future outbreaks, researchers need to know what studies have been done and what data they use. Mapping a large literature by hand is slow, while large language models alone can misstate or conflate sources. We present EpiLens, a lightweight, open-source retrieval-augmented generation package that applies a fixed retrieval template and output schema to every paper in a corpus and links outputs to source passages. We build a corpus of 2,297 papers from the sources of a multi-pathogen parameter study and evaluate EpiLens's paper-classification workflow against 214 manually labelled papers. Run locally with an open-weight model, so that licensed full texts never leave our hardware, the classifier is comparable to supervised baselines trained on the manual labels, and ahead of them on geography, despite the lack of task-specific training data. The study's early extraction sheets cite published sources for 327 estimates. We link each estimate to a cited document or to corpus studies included by a review. Then, we use the classifier's labels to compare the papers associated with different parameters and pathogens. EpiLens labels more papers linked to reproduction numbers than fatality risks as model-based (53% versus 15%, compared with 62% versus 24% in the manual labels). The gap is consistent with the frequent use of transmission models to estimate reproduction numbers and largely persists within pathogens.

cs.CL · cs.DL · physics.soc-ph

新论文 · v1arXiv:2611.04627

Evaluating LLMs as Judges of Counterspeech Quality Against Online Hate Speech

Johnny Alali, Nicolás Benjamín Ocampo, Davide Ceolin

原文摘要

Prolonged exposure to hate speech on social media can cause lasting psychological distress among users. Counterspeech has emerged as an effective, non-censorial mitigation strategy, but its deployment requires rigorous quality assessment. Manual assessment is unscalable due to the large volume of content on the Web and imposes severe emotional strain on human annotators. Large Language Models (LLMs) offer a scalable alternative, yet concerns remain regarding their reliability. In this paper, we investigate the extent to which LLMs can effectively perform this task. We test 12 open-weight LLMs of different sizes and families under zero-shot, few-shot, and chain-of-thought prompting, to approximate pointwise and pairwise human expert assessment of counterspeech quality across four dimensions: relevance, coherence, aggressiveness, and suitableness. We find that (i) LLMs show low agreement and moderate rank correlations with human experts: our best configuration, Qwen 2.5 72B on coherence, achieves up to a linear-weighted Cohen's $κ_l = 0.19$ and Spearman's $ρ= 0.42$ in pointwise evaluation and an unweighted Cohen's $κ= 0.25$ in pairwise evaluation. (ii) Converting pointwise scores into pairwise labels substantially improves human--LLM agreement for selected configurations. For same-instance coherence, Qwen 2.5 72B under zero-shot prompting achieves $κ= 0.35$ with converted labels, compared with $κ= 0.25$ when judging pairs directly, and Mixtral 8$\times$7B under few-shot prompting improves from $κ= 0.07$ to $κ= 0.34$. (iii) LLMs agree more with one another than with human experts, particularly in pairwise evaluation, where the strongest model pair, OLMo 2 32B and Llama 3.3 70B on coherence, reaches substantial agreement with $κ= 0.76$, whereas their agreement with the experts in the same configuration does not exceed $κ= 0.21$.

cs.CL

新论文 · v1arXiv:2611.04482

TEMPO: Time Series Understanding via Discrete Tokenization

Riccardo Maggioni

原文摘要

Time-series interpretation often requires both numerical observations and textual context. Integrating these inputs into a pretrained language model raises a representation question: how should a signal enter the model's token sequence? We present TEMPO, which extends an LLM's vocabulary with discrete codes from a context-aware Transformer tokenizer using finite scalar quantization. A frozen tokenizer maps four-step patches to a 625-code vocabulary; signal tokens, scale statistics, and text are then processed by the LLM's existing self-attention layers. Adaptation updates the new token embeddings and DoRA adapters. With a Qwen3-4B backbone and task-specific fine-tuning, TEMPO achieves macro-F1 scores of 89.0% on activity recognition, 85.0% on sleep staging, and 86.0% on bearing-fault diagnosis. Time-series question answering and qualitative bearing-diagnosis examples further illustrate how this interface combines signal representations with textual context. Together, these results demonstrate the utility of discrete vocabulary extension for time-series understanding.

cs.CL

版本更新 · v2arXiv:2611.00833

Fewer Tokens, More Answers: Verifiable-Reward Tuning of an Ornith-Style VLM for Multi-View Driving QA

T N Rickesh, Pranav Vinodh, Sowmya Kamath S

原文摘要

Reasoning-enabled model derivatives must turn intermediate deliberation into a usable answer within a deployment budget. We study that transition in an Ornith checkpoint adapted to multi-camera driving question answering with verifiable rewards. Paired archived outputs show that, on a selected 400-question development screen with a 256-token cap, GRPO increases answer extraction from 78.25% to 90.25% and accuracy from 55.50% to 63.75%. Of 39 wrong-to-correct transitions, 33 begin with an unextractable base response. The base's 87 unextractable responses all reach the token cap. Checkpoint-verified reconstruction on those same questions gives a much smaller first-pass accuracy gain at 512 tokens: +0.75 percentage points, versus +8.25 at 256. A later answer-cued continuation is analyzed separately; it helps the base more and narrows the adapter difference. On the full 2,511-question development set, this recovery-enabled pipeline reduces mean generated tokens from 108.6 to 47.1, while the paired accuracy interval includes zero gain. We frame these findings as adaptation of answer delivery under finite budgets, with explicit selection and provenance limits. Ornith's self-scaffolding lineage and newer reasoning-enabled derivatives such as Bonsai motivate the problem, but our measured results concern Ornith alone and do not establish a shared architectural mechanism or cross-model transfer.

cs.CV · cs.CL · cs.LG

新论文 · v1arXiv:2611.03617

Tracing the Scientific Discovery Capacity of LLMs to Their Training Data

Arda Uzunoglu, Mahler Revsine, Michael Schatz, Benjamin Van Durme, Mehrdad Farajtabar, Daniel Khashabi, Eric Nalisnick

原文摘要

Large language models (LLMs) show growing promise for scientific discovery, yet basic questions remain about what enables reliable discovery. How can we trace the scientific discovery capacity of LLMs to their training data? We study this question as one of missing probability mass and develop a theoretical account in which monofacts-the facts observed once during training-set the discovery capacity. In materials science and regulatory genomics, we directly manipulate monofact rates in training sets and find that more monofacts consistently yield more discoveries across two open-weight model families, sampling strategies, and reasoning budgets. Further experiments show the relationship cannot be explained by training-set diversity alone and can transfer across scientific domains when paired with in-domain data. Together, our results identify monofact rate in training data as a key factor governing discoveries in language models.

cs.CL · cs.AI

新论文 · v1arXiv:2611.03198

Written Once, Read Differently How Language Models Own Answers Steer Their Next Ones

Marcus Armstrong

原文摘要

When a language model answers a yes/no question, that answer changes how it answers a different question about the same entity. We study this \emph{self-commitment} in 22 open-weight models from four families (0.5B--72B parameters), fixing the model's first answer and measuring the shift in its second across 3,600 question pairs about 120 entities. Each answer writes a \emph{commitment direction} into the residual stream. The similarity of two questions' directions predicts their coupling on held-out entities ($ρ= 0.43$--$0.85$), and ablating a single direction shared across topics leaves at most 17\% of the effect in every model. How this direction is read is far less stable: the reading circuit spreads across more attention heads as models grow, instruction tuning makes more questions read it in the opposite sense, and in one model changing only the system prompt reverses the circuit's sign. Ablating the direction while sparing the attention sink removes most of the coupling at under 1.3\% perplexity cost in 18 of 19 instruction-tuned models, though it shifts standalone answers in two. Without the sink spared, the direction's alignment with the leading principal component of ordinary text predicts which models pay a high cost. Written once, the commitment is read differently across scale, training and prompt.

cs.CL · cs.AI

新论文 · v1arXiv:2611.01549

Revisiting LLM Homogeneity Through a Multidimensional Lens

Ziliang Qiu, Renfen Hu, Yi Ding, Ted Underwood

原文摘要

Concerns about LLM homogeneity are typically quantified with embedding similarity, which reduces two texts to a single score. We revisit homogeneity with 25 interpretable metrics spanning five dimensions: semantic, lexical, syntactic, structural, and stylistic & presentational. A human study shows that perceived similarity depends on both content (what responses say) and form (how they say it), yet embedding similarity correlates less with human meaning judgments than sentence-level matching ($ρ=0.417$ vs. $0.606$) and not significantly with form judgments ($ρ=0.140$). Across 18 contemporary LLMs, models say similar things in different words: they share global semantics and broad linguistic profiles but remain distinct in lexical characteristics, structural patterns, and interactional conventions. Across releases from 2024 to 2026, models become more like themselves but not more like one another: among significantly changing metrics, intra-model distances decrease for 17 of 22, and inter-model distances increase for 14 of 21. Across two post-training pipelines, model-specific conventions emerge during SFT and are selectively reshaped by preference optimization. These results suggest that homogeneity is multidimensional and evolving, and that embedding similarity alone can overstate cross-model homogeneity while concealing diverging intra- and inter-model trends.

cs.CL · cs.AI