arXiv subjects

Akhil Arora

Publications and source records attributed to Akhil Arora.

At least 19 recordsLinked to original sources

arXiv2026

The Working Set of a Coding Agent: Coherence Debt in Repository-Scale Tasks

Repository-scale coding requires an agent to keep tests, imports, configuration, and migration rules consistent within a bounded context window. We model this as reconstructing a coupled-fact graph: at each edit, a required fact comes from recent context or parametric memory, and the facts covered by neither form coherence debt. We supply and withhold each channel and inject faults across seven models and five harnesses. As expected, no model completes a task on an unseen API with both channels empty, and putting the facts in the prompt restores success. When a rename defeats what models memorized about a real library, all seven fail in the same place, passing and missing the same tests. Availability decides the outcome and distance does not: withholding a fact costs exactly the work it supports, and a supplied fact works as well far from the edit as next to it. Harnesses pay unequal prices for it: configurations that all pass every test differ more than tenfold in tokens consumed because they rebuild the same content at different rates, and spending more recovers nothing when facts are withheld. A missing fact produces wrong work rather than absent work: an agent asked to act acts, fabricating the file or guessing the value, so instruments built on reads look for a hole already filled. How often it says it is blocked instead is a property of the model, from every trial to none. Availability does not settle every edit: where standard and code disagree, agents follow the standard even when it prescribes the worse code, so a stale convention file costs more than no file. Because parametric memory substitutes for reading, on SWE-bench, where models likely know the repositories, reads no longer predict success. Harnesses should keep the facts an edit depends on available when the agent writes, and check that availability against what the agent produces rather than what it reads.

cs.SE↗

arXiv2026

Ghost Tool Calls: Issue-Time Privacy for Speculative Agent Tools

Tool-augmented language agents speculatively issue likely future tool calls to hide latency, but those calls leak inferred user intent to external services before the agent commits to the branch. Every external observer that received the call retains the disclosure after the agent abandons the branch. Timing is the issue, not authorization: no commit-time cleanup, read-only restriction, or access-control allow-list unsends what an observer already holds. We call these invocations ghost tool calls and propose Speculative Tool Privacy Contracts, a runtime abstraction that treats observation before commitment as a first-class effect, distinct from state mutation. We implement the contracts in a prototype runtime and evaluate twelve policies across three corpora. Speculative dispatch increases what an observer can infer about user intent; post-hoc filters, read-only restrictions, and access-control allow-lists leave that inference intact; only issue-time policies that change or suppress the speculative call's argument or destination projection before dispatch reduce it.

cs.CR↗

arXiv2026

Agentic Proving for Program Verification

Agentic systems have recently emerged as state-of-the-art approaches for automated theorem proving in formal mathematics. To assess how far these capabilities extend to program verification, we evaluate Claude Code in an agentic proving framework on CLEVER, a Lean 4 benchmark for verifiable code generation. Our results show that Claude generates arguably valid specifications for 98.8% of problems (with 81.3% also accepted by CLEVER's isomorphism-based scoring on the correct portion of the benchmark), certifies implementations against correct ground-truth specifications for 87.5% of problems, and reaches a 98.1% success rate on the end-to-end program generation and verification pipeline over entries with self-consistent premises. Across all stages, Claude further provides high-quality feedback on its own attempts (as confirmed under manual review), identifying underlying causes of failure and lingering bugs in the dataset. These findings highlight a growing mismatch between the difficulty of existing program verification benchmarks and the capabilities of modern agentic provers, and point to the need for more rigorous, bug-resilient evaluation methodologies, and in particular for alternatives to isomorphism-based scoring of generated specifications. More broadly, our results provide empirical evidence that tight compiler-in-the-loop agentic paradigms are currently the most effective approach for foundational program verification.

cs.AI↗

arXiv2026

MoBayes: A Modular Bayesian Framework for Separating Reasoning from Language in Conversational Clinical Decision Support

Large language models (LLMs) are increasingly used for conversational clinical decision support, yet they conflate next token prediction with probabilistic decision making. We argue that this conflation reflects an architectural limitation: such systems lack explicit posterior tracking, controllable abstention thresholds, and auditable reasoning chains. We introduce MoBayes, a Modular Bayesian dialogue framework that separates reasoning from language. The LLM acts only as a language interface, parsing patient conversation into structured observations, while a Bayesian module performs probabilistic inference over these observations to update posteriors, select follow-up questions via expected-information-gain and determine when to stop or defer through calibrated decision thresholds. This design enables explicit posterior tracking, controllable selective decision-making, and replaceable population-specific statistical backends without retraining the language model. Across empirical and LLM-generated knowledge bases, MoBayes outperforms standalone frontier LLM doctors, including matched model-family comparisons where inexpensive sensor models paired with MoBayes exceed larger autonomous models at lower cost. The advantage persists under adversarial patient communication styles and across varying diagnostic scenarios. These results suggest that reliable conversational clinical decision support systems should separate probabilistic reasoning from language generation rather than scaling model size alone. Code is available at https://anonymous.4open.science/r/MoBayes/

cs.LG↗

arXiv2026

Atomix: Timely, Transactional Tool Use for Reliable Agentic Workflows

LLM agents execute multi-step workflows that mutate external state through tools. Common orchestrators treat tool return as the settlement trigger, so faults, speculation, and concurrent agents can leave partial effects, losing-branch residue, stale writes, or irreversible sends. Correct settlement needs two facts that retries, checkpoint replay, locks, and compensation each conflate: which effects must settle together, and when earlier conflicting work is exhausted. Atomix makes this split explicit with progress-aware transactions. The runtime records reads and effects during execution, seals a transaction when its footprint is complete, and commits only after per-resource frontiers show that no earlier conflicting work can still arrive. Commit is final settlement: Atomix releases bufferable effects, accepts reversible external effects as final, and lets irreversible effects leave the gate. Abort suppresses unreleased effects and compensates externalized reversible effects where possible. On representative agent workloads, this composition improves clean recovery under injected faults, isolates contending and speculative work, and prevents correctly classified irreversible actions from leaking; microbenchmarks show microsecond-scale wrapper overhead relative to tool latency.

cs.LG↗

arXiv2026

DIVERGE: Diversity-Enhanced RAG for Open-Ended Information Seeking

Existing retrieval-augmented generation (RAG) systems often assume that each query has a single correct answer. This assumption overlooks open-ended information-seeking scenarios where multiple plausible answers are valuable, and where diversity is important for creativity, fairness, and inclusive access to information. We show that standard RAG systems fail to fully use diverse retrieved contexts: simply increasing retrieval diversity does not necessarily lead to diverse generations. To address this limitation, we propose Diverge, a plug-and-play agentic RAG framework that improves the diversity--quality trade-off through iterative, reflection-guided exploration of diverse viewpoints and diversity-aware retrieval support. We further introduce evaluation metrics for characterizing the diversity-quality trade-off in open-ended question answering. Experiments across multiple real-world datasets and backbone LLMs show that Diverge achieves the best trade-off among competitive baselines, increasing diversity by $\sim2\times$ without noticeable quality degradation. These results reveal a systematic limitation of current RAGs and show the value of explicit diversity modeling.

cs.CL↗

arXiv2025

ReasonBENCH: Benchmarking the (In)Stability of LLM Reasoning

Benchmark scores for LLM reasoning systems are reported as single numbers, yet the same model, strategy, and task can produce meaningfully different answers and costs across repeated executions, even under greedy decoding (T = 0). This variance is not a statistical nuisance: the highest-performing strategy wins only 77% of head-to-head runs against its nearest competitor, meaning a single observed score can silently misrank systems. We introduce ReasonBench, a benchmark suite recording 30 independent trials across 10 reasoning strategies, 12 models, and 6 tasks, treating quality and cost as distributions rather than point estimates. We find that this variance is structured rather than random: a two-component taxonomy -- Global Noise, capturing cross-benchmark unevenness, and Run Noise, capturing within-benchmark stochasticity -- reveals that strategy architecture predicts stability profiles, while models and strategies shift orthogonal aspects of the distribution. A hierarchical decomposition attributes three-quarters of score variance to benchmark, system, and item structure, with a persistent residual that single-run evaluation silently absorbs. Finally, cost and quality decouple asymmetrically: cheap methods are structurally immune to joint cost-quality failure, while expensive methods remain exposed regardless of their accuracy. These findings establish instability as an inherent property of reasoning systems and motivate distribution-aware evaluation as standard practice.

cs.AI↗

arXiv2025

Participatory AI: A Scandinavian Approach to Human-Centered AI

AI's transformative impact on work, education, and everyday life makes it as much a political artifact as a technological one. Current AI models are opaque, centralized, and overly generic. The algorithmic automation they provide threatens human agency and democratic values in both workplaces and daily life. To confront such challenges, we turn to Scandinavian Participatory Design (PD), which was devised in the 1970s to face a similar threat from mechanical automation. In the PD tradition, technology is seen not just as an artifact, but as a locus of democracy. Drawing from this tradition, we propose Participatory AI as a PD approach to human-centered AI that applies five PD principles to four design challenges for algorithmic automation. We use concrete case studies to illustrate how to treat AI models less as proprietary products and more as shared socio-technical systems that enhance rather than diminish human agency, human dignity, and human values.

cs.HC↗

arXiv2025

Quotegraph: A Social Network Extracted from Millions of News Quotations

We introduce Quotegraph, a novel large-scale social network derived from speaker-attributed quotations in English news articles published between 2008 and 2020. Quotegraph consists of 528 thousand unique nodes and 8.63 million directed edges, pointing from speakers to persons they mention. The nodes are linked to their corresponding items in Wikidata, thereby endowing the dataset with detailed biographic entity information, including nationality, gender, and political affiliation. Being derived from Quotebank, a massive corpus of quotations, relations in Quotegraph are additionally enriched with the information about the context in which they are featured. Each part of the network construction pipeline is language agnostic, enabling the construction of similar datasets based on non-English news corpora. We believe Quotegraph is a compelling resource for computational social scientists, complementary to online social networks, with the potential to yield novel insights into the behavior of public figures and how it is captured in the news.

cs.SI↗

arXiv2025

Web2Wiki: Characterizing Wikipedia Linking Across the Web

Wikipedia is one of the most visited websites globally, yet its role beyond its own platform remains largely unexplored. In this paper, we present the first large-scale analysis of how Wikipedia is referenced across the Web. Using a dataset from Common Crawl, we identify over 90 million Wikipedia links spanning 1.68% of Web domains and examine their distribution, context, and function. Our analysis of English Wikipedia reveals three key findings: (1) Wikipedia is most frequently cited by news and science websites for informational purposes, while commercial websites reference it less often. (2) The majority of Wikipedia links appear within the main content rather than in boilerplate or user-generated sections, highlighting their role in structured knowledge presentation. (3) Most links (95%) serve as explanatory references rather than as evidence or attribution, reinforcing Wikipedia's function as a background knowledge provider. While this study focuses on English Wikipedia, our publicly released Web2Wiki dataset includes links from multiple language editions, supporting future research on Wikipedia's global influence on the Web.

cs.SI↗

arXiv2025

Are Retrials All You Need? Enhancing Large Language Model Reasoning Without Verbalized Feedback

Recent advancements in large language models (LLMs) have catalyzed the development of general-purpose autonomous agents, demonstrating remarkable performance in complex reasoning tasks across various domains. This surge has spurred the evolution of a plethora of prompt-based reasoning frameworks. A recent focus has been on iterative reasoning strategies that refine outputs through self-evaluation and verbalized feedback. However, these strategies require additional computational complexity to enable models to recognize and correct their mistakes, leading to a significant increase in their cost. In this work, we introduce the concept of ``retrials without feedback'', an embarrassingly simple yet powerful mechanism for enhancing reasoning frameworks by allowing LLMs to retry problem-solving attempts upon identifying incorrect answers. Unlike conventional iterative refinement methods, our method does not require explicit self-reflection or verbalized feedback, simplifying the refinement process. Our findings indicate that simpler retrial-based approaches often outperform more sophisticated reasoning frameworks, suggesting that the benefits of complex methods may not always justify their computational costs. By challenging the prevailing assumption that more intricate reasoning strategies inherently lead to better performance, our work offers new insights into how simpler, more efficient approaches can achieve optimal results. So, are retrials all you need?

cs.CL↗

arXiv2025

Women, Infamous, and Exotic Beings: A Comparative Study of Honorific Usages in Wikipedia and LLMs for Bengali and Hindi

The obligatory use of third-person honorifics is a distinctive feature of several South Asian languages, encoding nuanced socio-pragmatic cues such as power, age, gender, fame, and social distance. In this work, (i) We present the first large-scale study of third-person honorific pronoun and verb usage across 10,000 Hindi and Bengali Wikipedia articles with annotations linked to key socio-demographic attributes of the subjects, including gender, age group, fame, and cultural origin. (ii) Our analysis uncovers systematic intra-language regularities but notable cross-linguistic differences: honorifics are more prevalent in Bengali than in Hindi, while non-honorifics dominate while referring to infamous, juvenile, and culturally exotic entities. Notably, in both languages, and more prominently in Hindi, men are more frequently addressed with honorifics than women. (iii) To examine whether large language models (LLMs) internalize similar socio-pragmatic norms, we probe six LLMs using controlled generation and translation tasks over 1,000 culturally balanced entities. We find that LLMs diverge from Wikipedia usage, exhibiting alternative preferences in honorific selection across tasks, languages, and socio-demographic attributes. These discrepancies highlight gaps in the socio-cultural alignment of LLMs and open new directions for studying how LLMs acquire, adapt, or distort social-linguistic norms. Our code and data are publicly available at https://github.com/souro/honorific-wiki-llm

cs.CL↗

arXiv2024

Entity Insertion in Multilingual Linked Corpora: The Case of Wikipedia

Links are a fundamental part of information networks, turning isolated pieces of knowledge into a network of information that is much richer than the sum of its parts. However, adding a new link to the network is not trivial: it requires not only the identification of a suitable pair of source and target entities but also the understanding of the content of the source to locate a suitable position for the link in the text. The latter problem has not been addressed effectively, particularly in the absence of text spans in the source that could serve as anchors to insert a link to the target entity. To bridge this gap, we introduce and operationalize the task of entity insertion in information networks. Focusing on the case of Wikipedia, we empirically show that this problem is, both, relevant and challenging for editors. We compile a benchmark dataset in 105 languages and develop a framework for entity insertion called LocEI (Localized Entity Insertion) and its multilingual variant XLocEI. We show that XLocEI outperforms all baseline models (including state-of-the-art prompt-based ranking with LLMs such as GPT-4) and that it can be applied in a zero-shot manner on languages not seen during training with minimal performance drop. These findings are important for applying entity insertion models in practice, e.g., to support editors in adding links across the more than 300 language versions of Wikipedia.

cs.CL↗

arXiv2024

Fleet of Agents: Coordinated Problem Solving with Large Language Models

While numerous frameworks have been developed to enhance the reasoning abilities of large language models (LLMs), there is a scarcity of methods that effectively balance the trade-off between cost and quality. In this paper, we introduce Fleet of Agents (FoA), a novel and intuitive yet principled framework utilizing LLMs as agents to navigate through dynamic tree searches, employing a genetic-type particle filtering approach. FoA spawns a multitude of agents, each exploring the search space autonomously, followed by a selection phase where resampling based on a heuristic value function optimizes the balance between exploration and exploitation. This mechanism enables dynamic branching, adapting the exploration strategy based on discovered solutions. We conduct extensive experiments on three benchmark tasks, ``Game of 24'', ``Mini-Crosswords'', and ``WebShop'', utilizing four different LLMs, ``GPT-3.5'', ``GPT-4'', ``LLaMA3.2-11B'', and ``LLaMA3.2-90B''. On average across all tasks and LLMs, FoA obtains a quality improvement of ~5% while requiring only ~40% of the cost of previous SOTA methods. Notably, our analyses reveal that (1) FoA achieves the best cost-quality trade-off among all benchmarked methods and (2) FoA + LLaMA3.2-11B surpasses the Llama3.2-90B model. FoA is publicly available at https://github.com/au-clan/FoA.

cs.CL↗

arXiv2023

Orphan Articles: The Dark Matter of Wikipedia

With 60M articles in more than 300 language versions, Wikipedia is the largest platform for open and freely accessible knowledge. While the available content has been growing continuously at a rate of around 200K new articles each month, very little attention has been paid to the accessibility of the content. One crucial aspect of accessibility is the integration of hyperlinks into the network so the articles are visible to readers navigating Wikipedia. In order to understand this phenomenon, we conduct the first systematic study of orphan articles, which are articles without any incoming links from other Wikipedia articles, across 319 different language versions of Wikipedia. We find that a surprisingly large extent of content, roughly 15\% (8.8M) of all articles, is de facto invisible to readers navigating Wikipedia, and thus, rightfully term orphan articles as the dark matter of Wikipedia. We also provide causal evidence through a quasi-experiment that adding new incoming links to orphans (de-orphanization) leads to a statistically significant increase of their visibility in terms of the number of pageviews. We further highlight the challenges faced by editors for de-orphanizing articles, demonstrate the need to support them in addressing this issue, and provide potential solutions for developing automated tools based on cross-lingual approaches. Overall, our work not only unravels a key limitation in the link structure of Wikipedia and quantitatively assesses its impact, but also provides a new perspective on the challenges of maintenance associated with content creation at scale in Wikipedia.

cs.SI↗

arXiv2023

Generating Faithful Synthetic Data with Large Language Models: A Case Study in Computational Social Science

Large Language Models (LLMs) have democratized synthetic data generation, which in turn has the potential to simplify and broaden a wide gamut of NLP tasks. Here, we tackle a pervasive problem in synthetic data generation: its generative distribution often differs from the distribution of real-world data researchers care about (in other words, it is unfaithful). In a case study on sarcasm detection, we study three strategies to increase the faithfulness of synthetic data: grounding, filtering, and taxonomy-based generation. We evaluate these strategies using the performance of classifiers trained with generated synthetic data on real-world data. While all three strategies improve the performance of classifiers, we find that grounding works best for the task at hand. As synthetic data generation plays an ever-increasing role in NLP research, we expect this work to be a stepping stone in improving its utility. We conclude this paper with some recommendations on how to generate high(er)-fidelity synthetic data for specific tasks.

cs.CL↗

arXiv2022

Quote Erat Demonstrandum: A Web Interface for Exploring the Quotebank Corpus

The use of attributed quotes is the most direct and least filtered pathway of information propagation in news. Consequently, quotes play a central role in the conception, reception, and analysis of news stories. Since quotes provide a more direct window into a speaker's mind than regular reporting, they are a valuable resource for journalists and researchers alike. While substantial research efforts have been devoted to methods for the automated extraction of quotes from news and their attribution to speakers, few comprehensive corpora of attributed quotes from contemporary sources are available to the public. Here, we present an adaptive web interface for searching Quotebank, a massive collection of quotes from the news, which we make available at https://quotebank.dlab.tools.

cs.IR↗

arXiv2022

Strong Heuristics for Named Entity Linking

Named entity linking (NEL) in news is a challenging endeavour due to the frequency of unseen and emerging entities, which necessitates the use of unsupervised or zero-shot methods. However, such methods tend to come with caveats, such as no integration of suitable knowledge bases (like Wikidata) for emerging entities, a lack of scalability, and poor interpretability. Here, we consider person disambiguation in Quotebank, a massive corpus of speaker-attributed quotations from the news, and investigate the suitability of intuitive, lightweight, and scalable heuristics for NEL in web-scale corpora. Our best performing heuristic disambiguates 94% and 63% of the mentions on Quotebank and the AIDA-CoNLL benchmark, respectively. Additionally, the proposed heuristics compare favourably to the state-of-the-art unsupervised and zero-shot methods, Eigenthemes and mGENRE, respectively, thereby serving as strong baselines for unsupervised and zero-shot entity linking.

cs.CL↗