arXiv subjects

Deepali Mishra

Publications and source records attributed to Deepali Mishra.

5 recordsLinked to original sources

arXiv2026

SycoPhantasy: Quantifying Sycophancy and Hallucination in Small Open Weight VLMs for Vision-Language Scoring of Fantasy Characters

Vision-language models (VLMs) are increasingly deployed as evaluators in tasks requiring nuanced image understanding, yet their reliability in scoring alignment between images and text descriptions remains underexplored. We investigate whether small, open-weight VLMs exhibit \emph{sycophantic} behavior when evaluating image-text alignment: assigning high scores without grounding their judgments in visual evidence. To quantify this phenomenon, we introduce the \emph{Bluffing Coefficient} (\bc), a metric that measures the mismatch between a model's score and its evidence recall. We evaluate six open-weight VLMs ranging from 450M to 8B parameters on a benchmark of 173,810 AI-generated character portraits paired with detailed textual descriptions. Our analysis reveals a significant inverse correlation between model size and sycophancy rate ($r = -0.96$, $p = 0.002$), with smaller models exhibiting substantially higher rates of unjustified high scores. The smallest model tested (LFM2-VL, 450M) produced sycophantic evaluations in 22.3\% of cases, compared to 6.0\% for the largest (LLaVA-1.6, 7B). These findings have direct implications for the deployment of small, open-weight VLMs as automated evaluators within attribute-rich, synthetic image evaluation tasks, where the gap between assigned scores and cited visual evidence is both measurable and consequential.

cs.CV↗

arXiv2026

Too Nice to Tell the Truth: Quantifying Agreeableness-Driven Sycophancy in Role-Playing Language Models

Large language models increasingly serve as conversational agents that adopt personas and role-play characters at user request. This capability, while valuable, raises concerns about sycophancy: the tendency to provide responses that validate users rather than prioritize factual accuracy. While prior work has established that sycophancy poses risks to AI safety and alignment, the relationship between specific personality traits of adopted personas and the degree of sycophantic behavior remains unexplored. We present a systematic investigation of how persona agreeableness influences sycophancy across 13 small, open-weight language models ranging from 0.6B to 20B parameters. We develop a benchmark comprising 275 personas evaluated on NEO-IPIP agreeableness subscales and expose each persona to 4,950 sycophancy-eliciting prompts spanning 33 topic categories. Our analysis reveals that 9 of 13 models exhibit statistically significant positive correlations between persona agreeableness and sycophancy rates, with Pearson correlations reaching $r = 0.87$ and effect sizes as large as Cohen's $d = 2.33$. These findings demonstrate that agreeableness functions as a reliable predictor of persona-induced sycophancy, with direct implications for the deployment of role-playing AI systems and the development of alignment strategies that account for personality-mediated deceptive behaviors.

cs.CL↗

arXiv2025

Barriers in Integrating Medical Visual Question Answering into Radiology Workflows: A Scoping Review and Clinicians' Insights

Medical Visual Question Answering (MedVQA) is a promising tool to assist radiologists by automating medical image interpretation through question answering. Despite advances in models and datasets, MedVQA's integration into clinical workflows remains limited. This study systematically reviews 68 publications (2018-2024) and surveys 50 clinicians from India and Thailand to examine MedVQA's practical utility, challenges, and gaps. Following the Arksey and O'Malley scoping review framework, we used a two-pronged approach: (1) reviewing studies to identify key concepts, advancements, and research gaps in radiology workflows, and (2) surveying clinicians to capture their perspectives on MedVQA's clinical relevance. Our review reveals that nearly 60% of QA pairs are non-diagnostic and lack clinical relevance. Most datasets and models do not support multi-view, multi-resolution imaging, EHR integration, or domain knowledge, features essential for clinical diagnosis. Furthermore, there is a clear mismatch between current evaluation metrics and clinical needs. The clinician survey confirms this disconnect: only 29.8% consider MedVQA systems highly useful. Key concerns include the absence of patient history or domain knowledge (87.2%), preference for manually curated datasets (51.1%), and the need for multi-view image support (78.7%). Additionally, 66% favor models focused on specific anatomical regions, and 89.4% prefer dialogue-based interactive systems. While MedVQA shows strong potential, challenges such as limited multimodal analysis, lack of patient context, and misaligned evaluation approaches must be addressed for effective clinical integration.

cs.CL↗

arXiv2019

Generalised Garfinkle-Vachaspati Transform With Dilaton

The generalised Garfinkle-Vachaspati transform (GGV) introduced in our previous work (arXiv:1808.04981) is a powerful technique to add hair modes to a class of known D1-D5 solutions. The richest and most interesting examples of D1-D5 geometries involve the dilaton and this calls for an extension of the procedure to accommodate for the presence of the dilaton field. In this paper we present such extended version of the GGV transform. We explore this generalisation in ten-dimensions with R-R or NS-NS 2-form field and relate the two set-ups via S-duality. In the context of the D1-D5 system, this generalisation allows us to add travelling wave deformations on solutions beyond minimal six-dimensional supergravity lifted to IIB supergravity. We work out travelling wave deformations involving the torus directions on a class of supersymmetric D1-D5-P geometries. We also explore applications of our technique to the F1-P system.

hep-th↗

arXiv2018

A Generalised Garfinkle-Vachaspati Transform

The Garfinkle-Vachaspati transform is a deformation of a metric in terms of a null, hypersurface orthogonal, Killing vector $k^μ$. We explore a generalisation of this deformation in type IIB supergravity taking motivation from certain studies of the D1-D5 system. We consider solutions of minimal six-dimensional supergravity admitting null Killing vector $k^μ$ trivially lifted to type IIB supergravity by the addition of four-torus directions. The torus directions provide covariantly constant spacelike vectors $l^μ$. We show that the original solution can be deformed as $g_{μν} \to g_{μν} + 2 Φk_{(μ}l_{ν)}, C_{μν} \to C_{μν} - 2 Φk_{[μ}l_{ν]}$, provided the two-form supporting the original spacetime satisfies $i_k (dC) = - d k$, and where $Φ$ satisfies the equation of a minimal massless scalar field on the original spacetime. We show that the condition $i_k (dC) = - d k$ is satisfied by all supersymmetric solutions admitting null Killing vector. Hence all supersymmetric solutions of minimal six-dimensional supergravity can be deformed via this method. As an example of our approach, we work out the deformation on a class of D1-D5-P geometries with orbifolds. We show that the deformed spacetimes are smooth and identify their CFT description. Using Bena-Warner formalism, we also express the deformed solutions in other duality frames.

hep-th↗