Tootfinder

Deep opacity and AI: A threat to XAI and to privacy protection mechanisms
Vincent C. M\"uller
https://arxiv.org/abs/2509.08835 https://arxiv.org/pdf/2…

Deep opacity and AI: A threat to XAI and to privacy protection mechanisms
It is known that big data analytics and AI pose a threat to privacy, and that some of this is due to some kind of "black box problem" in AI. I explain how this becomes a problem in the context of justification for judgments and actions. Furthermore, I suggest distinguishing three kinds of opacity: 1) the subjects do not know what the system does ("shallow opacity"), 2) the analysts do not know what the system does ("standard black box opacity"), or 3) the analysts cannot possibly know what the …

MRI-Based Brain Tumor Detection through an Explainable EfficientNetV2 and MLP-Mixer-Attention Architecture
Mustafa Yurdakul, \c{S}akir Ta\c{s}demir
https://arxiv.org/abs/2509.06713

MRI-Based Brain Tumor Detection through an Explainable EfficientNetV2 and MLP-Mixer-Attention Architecture
Brain tumors are serious health problems that require early diagnosis due to their high mortality rates. Diagnosing tumors by examining Magnetic Resonance Imaging (MRI) images is a process that requires expertise and is prone to error. Therefore, the need for automated diagnosis systems is increasing day by day. In this context, a robust and explainable Deep Learning (DL) model for the classification of brain tumors is proposed. In this study, a publicly available Figshare dataset containing 3,…

Crosslisted article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[3/6]:
- In-Context Policy Adaptation via Cross-Domain Skill Diffusion
Minjong Yoo, Woo Kyung Kim, Honguk Woo

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[3/8]:
- TACO: Enhancing Multimodal In-context Learning via Task Mapping-Guided Sequence Configuration
Yanshu Li, Jianjiang Yang, Tian Yun, Pinyuan Feng, Jinfa Huang, Ruixiang Tang

If you think there might be an AI bubble, but don't worry, even if it pops, how bad can it be?
Here's one number: seven tech companies (NVIDIA, Microsoft, Google, Apple, Meta, Tesla, Amazon) are worth ~20 trillion, ⇓ of total US stock market
For context: total subprime mortgages in 2008 were 1.3 trillion, total US mortgage debt 10 trillion
You have a 401k or index fonds? Your money is likely tied with the AI market. If you want that, that's fine. If you're …

The Masses of Fermions in the context of the Supersymmetric $SU(3)_{C}\times SU(3)_{L}\times U(1)_{N}$ Model
M. C. Rodriguez
https://arxiv.org/abs/2508.11456 https://

The Masses of Fermions in the context of the Supersymmetric $SU(3)_{C}\times SU(3)_{L}\times U(1)_{N}$ Model
We will study in details the masses spectrum of fermions of the Minimal Supersymmetric $SU(3)_{C}\times SU(3)_{L}\times U(1)_{N}$ Model. We will consider the case in which all the usual neutral scalars fields of this model obtain vacuum expectation values, in both case when we have $R$-parity conservation, respect the invariance of the quantum number, ${\cal F}\equiv B+L$, where $B$ is the baryon number while $L$ is the total lepton number, and the oppposite ones, it means $R$-parity violation.…

Windows 11 Intlune devices disconnecting from Entra ID - devices no longer Entra Joined after reboot
https://www.reddit.com/r/Intune/comments/1od595p/comment/nktjcww/?context=3

Formal Models and Convergence Analysis for Context-Aware Security Verification
Ayush Chaudhary
https://arxiv.org/abs/2510.12440 https://arxiv.org/pdf/2510.…

Formal Models and Convergence Analysis for Context-Aware Security Verification
We present a formal framework for context-aware security verification that establishes provable guarantees for ML-enhanced adaptive systems. We introduce context-completeness - a new security property - and prove: (1) sample complexity bounds showing when adaptive verification succeeds, (2) information-theoretic limits relating context richness to detection capability, (3) convergence guarantees for ML-based payload generators, and (4) compositional soundness bounds. We further provide a formal…

Combinatorics of monoidal actions in Lie-algebraic context
Volodymyr Mazorchuk, Xiaoyu Zhu
https://arxiv.org/abs/2509.01404 https://arxiv.org/pdf/2509.0140…

Combinatorics of monoidal actions in Lie-algebraic context
This paper is, essentially, a survey related to the problem of understanding the combinatorics of the action of the monoidal category of finite dimensional modules over a simple finite dimensional Lie algebra on various categories of Lie algebra modules. A special attention is payed to the Lie algebras $\mathfrak{sl}_2$ and $\mathfrak{sl}_3$. A few new general results are collected at the end.

Finding My Way: Influence of Different Audio Augmented Reality Navigation Cues on User Experience and Subjective Usefulness
Sina Hinzmann, Francesco Vona, Juliane Henning, Mohamed Amer, Omar Abdellatif, Tanja Kojic, Jan-Niklas Voigt-Antons
https://arxiv.org/abs/2509.03199

Finding My Way: Influence of Different Audio Augmented Reality Navigation Cues on User Experience and Subjective Usefulness
As augmented reality (AR) becomes increasingly prevalent in mobile and context-aware applications, the role of auditory cues in guiding users through physical environments is becoming critical. This study investigates the effectiveness and user experience of various categories of audio cues, including fully non-verbal sounds and speech-derived Spearcons, during outdoor navigation tasks using the Meta Quest 3 headset. Twenty participants navigated five outdoor routes using audio-only cue types: …

Next highest weight and other lower $SU(3)$ irreducible representations with proxy-$SU(4)$ symmetry for nuclei with $32 \le \mbox{Z,N} \le 46$
V. K. B. Kota
https://arxiv.org/abs/2510.00800

Next highest weight and other lower $SU(3)$ irreducible representations with proxy-$SU(4)$ symmetry for nuclei with $32 \le \mbox{Z,N} \le 46$
In the applications of proxy-SU(3) model in the context of determining $(β,γ)$ values for nuclei across the periodic table, for understanding the preponderance of triaxial shapes in nuclei with $Z \ge 30$, it is seen that one needs not only the highest weight (hw) or leading $SU(3)$ irreducible representation (irrep) $(λ_H, μ_H)$ but also the lower $SU(3)$ irreps $(λ,μ)$ such that $2λ+ μ=2λ_H + μ_H-3r$ with $r=0,1$ and $2$ [Bonatsos et al., Symmetry {\bf 16}, 1625 (2024)]. These give …

A JWST/MIRI view of k Andromedae b: Refining its mass, age, and physical parameters
N. Godoy, E. Choquet, E. Serabyn, M. Malin, P. Tremblin, C. Danielski, P. O. Lagage, A. Boccaletti, B. Charnay, M. E. Ressler
https://arxiv.org/abs/2509.03624

A JWST/MIRI view of k Andromedae b: Refining its mass, age, and physical parameters
Context. kAndb is a substellar companion near the planet-brown dwarf boundary, orbiting a B9IV star at 50-100 au. Estimates of its age and mass vary, fueling a decade-long debate. Atmospheric parameters (Teff 1650-2050 K, log(g) 3.5-5.5) remain poorly constrained due to model differences and heterogeneous datasets. Aims. We refine the characterization of kAndb using mid-infrared data to capture its bolometric emission. Combined with NIR measurements, we constrain Teff, log(g), and radius to red…

${C}^{3}$-GS: Learning Context-aware, Cross-dimension, Cross-scale Feature for Generalizable Gaussian Splatting
Yuxi Hu, Jun Zhang, Kuangyi Chen, Zhe Zhang, Friedrich Fraundorfer
https://arxiv.org/abs/2508.20754

${C}^{3}$-GS: Learning Context-aware, Cross-dimension, Cross-scale Feature for Generalizable Gaussian Splatting
Generalizable Gaussian Splatting aims to synthesize novel views for unseen scenes without per-scene optimization. In particular, recent advancements utilize feed-forward networks to predict per-pixel Gaussian parameters, enabling high-quality synthesis from sparse input views. However, existing approaches fall short in encoding discriminative, multi-view consistent features for Gaussian predictions, which struggle to construct accurate geometry with sparse views. To address this, we propose $\m…

Limiting the Parameter Space for Unstable eV-scale Neutrinos Using IceCube Data
Abbasi, Ackermann, Adams, Agarwalla, Aguilar, Ahlers, Alameddine, Ali, Amin, Andeen, Arg\"uelles, Ashida, Athanasiadou, Axani, Babu, Bai, Baines-Holmes, V., Barwick, Bash, Basu, Bay, Beatty, Tjus, Behrens, Beise, Bellenghi, Benkel, BenZvi, Berley, Bernardini, Besson, Blaufuss, Bloom, Blot, Bodo, Bontempo, Motzkin, Meneguolo, B\"oser, Botner, B\"ottcher, Braun, Brinson, Brisson-Tsavoussis, Bur…

Limiting the Parameter Space for Unstable eV-scale Neutrinos Using IceCube Data
This Letter extends a recent IceCube sterile neutrino search to include unstable sterile neutrinos within the context of a model termed 3+1+Decay, which expands upon the 3+1 model by introducing sterile neutrino decay to invisible particles with coupling constant $g^2$. The model is attractive since it reduces tension between oscillation experiments within the global fits and with constraints that come from cosmological observables. The analysis uses 10.7 years of up-going muon neutrino data wi…

Studying the black widow pulsars PSR J0312$-$0921 and PSR J1627$ $3219 in the optical and X-rays
A. V. Bobakov, A. Kirichenko, S. V. Zharikov, D. A. Zyuzin, A. V. Karpova, Yu. A. Shibanov, T. Begari
https://arxiv.org/abs/2509.01488

Studying the black widow pulsars PSR J0312$-$0921 and PSR J1627$+$3219 in the optical and X-rays
Context. PSR J0312$-$0921 and PSR J1627$+$3219 are black widow pulsars with orbital periods of 2.34 and 3.98 hours. They were recently detected in the radio and $γ$-rays. Aims. Our goals are to estimate the fundamental parameters of both binary systems and their components. Methods. We performed first phase-resolved multi-band photometry of both objects with the 10.4-m Gran Telescopio Canarias and fitted the obtained light curves with a model assuming direct heating of the companion by the pul…

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[2/3]:
- Context Reasoner: Incentivizing Reasoning Capability for Contextualized Privacy and Safety Compli...
Hu, Li, Jing, Hu, Zeng, Han, Xu, Chu, Hu, Song

@… thanks, I wonder whether Ventoy is compatible with 15.0.
https://www.reddit.com/r/freebsd/comments/1nh3z73/comment/nfe4kd9/?contex…

Replaced article(s) found for nlin.CD. https://arxiv.org/list/nlin.CD/new
[1/1]:
- Context parroting: A simple but tough-to-beat baseline for foundation models in scientific machin...
Yuanzhao Zhang, William Gilpin

The Impact of Axion-Like Particles on Late Stellar Evolution From Intermediate-Mass Stars to core-collapse Supernova Progenitors
Inmacolata Dom\'inguez, Oscar Straniero, Luciano Piersanti, Maurizio Giannotti, Alessandro Mirizzi
https://arxiv.org/abs/2508.17779

The Impact of Axion-Like Particles on Late Stellar Evolution From Intermediate-Mass Stars to core-collapse Supernova Progenitors
Context. Stars with masses ranging from 3 to 11 M_\odot exhibit multiple evolutionary paths. Less massive stars in this range conclude their evolution as carbon-oxygen (CO) white dwarfs. However, those that achieve carbon ignition before the pressure by degenerate electron halts the core contraction may either form massive CONe/ONe white dwarfs, or undergo an electron-capture supernova, or photo-disintegrate neon and proceed with further thermonuclear burning, ultimately leading to the formatio…

GPT-OSS-20B: A Comprehensive Deployment-Centric Analysis of OpenAI's Open-Weight Mixture of Experts Model
Deepak Kumar, Divakar Yadav, Yash Patel
https://arxiv.org/abs/2508.16700

GPT-OSS-20B: A Comprehensive Deployment-Centric Analysis of OpenAI's Open-Weight Mixture of Experts Model
We present a single-GPU (H100, bf16) evaluation of GPT-OSS-20B (Mixture-of-Experts; 20.9B total, approx. 3.61B active) against dense baselines Qwen3-32B and Yi-34B across multiple dimensions. We measure true time-to-first-token (TTFT), full-decode throughput (TPOT), end-to-end latency percentiles, peak VRAM with past key values (PKV) held, and energy via a consistent nvidia-smi-based sampler. At a 2048-token context with 64-token decode, GPT-OSS-20B delivers higher decode throughput and tokens …

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[3/4]:
- Dynamic Context Compression for Efficient RAG
Shuyu Guo, Zhaochun Ren
https://

On the analytic rank of the twin prime elliptic curve $y^2=x(x-2)(x-p)$
Kirti Joshi
https://arxiv.org/abs/2508.12340 https://arxiv.org/pdf/2508.12340

On the analytic rank of the twin prime elliptic curve $y^2=x(x-2)(x-p)$
Let $p\geq 7$ and suppose $(p,p-2)$ are twin prime numbers, in [Hatley, 2009], the elliptic curve $E_p:y^2=x(x-2)(x-p)$ was considered in the context of a conjecture by Jason Beers about the Mordell-Weil ranks of $E_p/\mathbb{Q}$. I show that for $p\equiv 3,5\bmod 8$, the analytic rank of $E_p$ is at least one (Theorem 1.1.2) in line with Beers' predictions. This is done by finding a formula (Theorem 4.1.1) for the global root number of $E_p$ for all twin prime pairs. I also show that Beers' co…

Restricted Jacobi permutations
Alyssa G. Henke, Kyle R. Hoffman, Derek H. Stephens, Yongwei Yuan, Yan Zhuang
https://arxiv.org/abs/2509.11494 https://arxiv…

Restricted Jacobi permutations
Jacobi permutations, introduced by Viennot in the context of Jacobi elliptic functions, are counted by the Euler numbers $E_{n}$ appearing in the series expansion $\sec x+\tan x=\sum_{n=0}^{\infty}E_{n}x^{n}/n!$. We conduct a systematic study of pattern avoidance in Jacobi permutations, achieving a complete enumeration of Jacobi permutations avoiding a prescribed set of length 3 patterns. In the case of a single pattern restriction, we obtain refined enumerations with respect to several permuta…

Theory space and stability analysis of General Relativistic cosmological solutions in modified gravity
Saikat Chakraborty, Piyabut Burikham
https://arxiv.org/abs/2509.15762 http…

Theory space and stability analysis of General Relativistic cosmological solutions in modified gravity
Some aspects of two General Relativistic cosmological solutions, an exact $Λ$CDM-like cosmological solution $j=1$ ($j$ is cosmographic jerk parameter), and a specifically designed toy cosmological solution $j=1+3\varepsilon(q-1/2)$ ($q$ is cosmographic deceleration parameter, $0<|\varepsilon|<1$) that is capable of accommodating a phantom crossing scenario as suggested by DESI DR2, are studied within the context of $f(R)$ gravity, by portraying them as a \emph{flow} in the 2-dimensio…

Early Approaches to Adversarial Fine-Tuning for Prompt Injection Defense: A 2022 Study of GPT-3 and Contemporary Models
Gustavo Sandoval, Denys Fenchenko, Junyao Chen
https://arxiv.org/abs/2509.14271

Early Approaches to Adversarial Fine-Tuning for Prompt Injection Defense: A 2022 Study of GPT-3 and Contemporary Models
This paper documents early research conducted in 2022 on defending against prompt injection attacks in large language models, providing historical context for the evolution of this critical security domain. This research focuses on two adversarial attacks against Large Language Models (LLMs): prompt injection and goal hijacking. We examine how to construct these attacks, test them on various LLMs, and compare their effectiveness. We propose and evaluate a novel defense technique called Adversar…

Mapping Toxic Comments Across Demographics: A Dataset from German Public Broadcasting
Jan Fillies, Michael Peter Hoffmann, Rebecca Reichel, Roman Salzwedel, Sven Bodemer, Adrian Paschke
https://arxiv.org/abs/2508.21084

Mapping Toxic Comments Across Demographics: A Dataset from German Public Broadcasting
A lack of demographic context in existing toxic speech datasets limits our understanding of how different age groups communicate online. In collaboration with funk, a German public service content network, this research introduces the first large-scale German dataset annotated for toxicity and enriched with platform-provided age estimates. The dataset includes 3,024 human-annotated and 30,024 LLM-annotated anonymized comments from Instagram, TikTok, and YouTube. To ensure relevance, comments we…

On the boundary Carrollian conformal algebra
Lucas Buzaglo, Xiao He, Tuan Anh Pham, Haijun Tan, Girish S Vishwa, Kaiming Zhao
https://arxiv.org/abs/2508.21603 https://

On the boundary Carrollian conformal algebra
We initiate the mathematical study of the boundary Carrollian conformal algebra (BCCA), an infinite-dimensional Lie algebra recently discovered in the context of Carrollian physics. The BCCA is an intriguing object from both physical and mathematical perspectives, since it is a filtered but not graded Lie algebra. In this paper, we first construct some modules for the BCCA and one of its subalgebras, which we call $\mathcal{O}$, by restriction of well-known modules of the BMS$_3$ and Witt algeb…

Reionization in HESTIA: Studying reionization in the LG through zoom simulations
David Attard, Luke Conaboy, Noam Libeskind, Sergey Pillipenko, Keri Dixon, Ilian T. Iliev
https://arxiv.org/abs/2509.10133

Reionization in HESTIA: Studying reionization in the LG through zoom simulations
While cosmic reionization has been broadly constrained by global observables, the interplay between internal sources (Milky Way, M31, and their satellites) and external ionization fronts remains poorly understood in a realistic Local Group (LG) context. To address this issue, we perform radiative transfer post-processing on the original HESTIA LG constrained simulation. We calibrate our source models using a uniform 1024^3 particle, dark-matter only, HESTIA simulation coupled with a subgrid col…

The higher spin $\Pi$-operator in Clifford analysis
Wanqing Cheng, Chao Ding
https://arxiv.org/abs/2508.11271 https://arxiv.org/pdf/2508.11271

The higher spin $Π$-operator in Clifford analysis
Rarita-Schwinger fields are solutions to the relativistic field equation of spin-$3/2$ fermions in four dimensional flat spacetime, which are important in supergravity and superstring theories. Bure\v s et al. generalized it to arbitrary spin $k/2$ in 2002 in the context of Clifford algebras. In this article, we introduce the higher spin $Π$-operator related to the Rarita-Schwinger operator. Further, we investigate norm estimates, mapping properties and the adjoint operator of the higher spin …

The Interstellar Medium in IZw18 seen with JWST/MIRI: I. Highly Ionized Gas
L. K. Hunt, A. Aloisi, M. G. Navarro, R. J. Rickards Vaught, B. T. Draine, A. Adamo, F. Annibali, D. Calzetti, S. Hernandez, B. L. James, M. Mingozzi, R. Schneider, M. Tosi, B. Brandl, M. G. del Valle-Espinosa, F. Donnan, A. S. Hirschauer, M. Meixner, D. Rigopoulou, C. T. Richardson, J. M. Levanti, A. R. Basu-Zych

The Interstellar Medium in IZw18 seen with JWST/MIRI: I. Highly Ionized Gas
We present JWST/MIRI spectra from the Medium-Resolution Spectrometer of IZw18, a nearby dwarf galaxy with a metallicity of $\sim$3% Solar. Its proximity enables a detailed study of highly ionized gas that can be interpreted in the context of newly discovered high-redshift dwarf galaxies. We derive aperture spectra centered on eleven regions of interest; the spectra show very low extinction, A_V $\lesssim 0.1$, consistent with optical determinations. The gas is highly ionized; we have detected 1…

LLMs in the SOC: An Empirical Study of Human-AI Collaboration in Security Operations Centres
Ronal Singh, Shahroz Tariq, Fatemeh Jalalvand, Mohan Baruwal Chhetri, Surya Nepal, Cecile Paris, Martin Lochner
https://arxiv.org/abs/2508.18947

LLMs in the SOC: An Empirical Study of Human-AI Collaboration in Security Operations Centres
The integration of Large Language Models (LLMs) into Security Operations Centres (SOCs) presents a transformative, yet still evolving, opportunity to reduce analyst workload through human-AI collaboration. However, their real-world application in SOCs remains underexplored. To address this gap, we present a longitudinal study of 3,090 analyst queries from 45 SOC analysts over 10 months. Our analysis reveals that analysts use LLMs as on-demand aids for sensemaking and context-building, rather th…

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[3/6]:
- A Layered Multi-Expert Framework for Long-Context Mental Health Assessments
Jinwen Tang, Qiming Guo, Wenbo Sun, Yi Shang

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[3/5]:
- ILRe: Intermediate Layer Retrieval for Context Compression in Causal Language Models
Manlai Liang, Mandi Liu, Jiangzhou Ji, Huaijun Li, Haobo Yang, Yaohan He, Jinlong Li

Local existence theory for a class of CMC gauges for the Einstein-non-linear scalar field equations
Hans Ringstr\"om
https://arxiv.org/abs/2509.14110 https://

Local existence theory for a class of CMC gauges for the Einstein-non-linear scalar field equations
The purpose of this article is to develop a local existence theory for a class of CMC gauges for the Einstein-non-linear scalar field equations. We do so in the context of closed and parallelisable initial manifolds. The assumption that the initial manifold is parallelisable is not a topological restriction in the case of closed oriented 3-manifolds, but it is a restriction in higher dimensions. The results include local existence, uniqueness, a continuation criterion and Cauchy stability. Prev…

The Few-shot Dilemma: Over-prompting Large Language Models
Yongjian Tang, Doruk Tuncel, Christian Koerner, Thomas Runkler
https://arxiv.org/abs/2509.13196 https://

The Few-shot Dilemma: Over-prompting Large Language Models
Over-prompting, a phenomenon where excessive examples in prompts lead to diminished performance in Large Language Models (LLMs), challenges the conventional wisdom about in-context few-shot learning. To investigate this few-shot dilemma, we outline a prompting framework that leverages three standard few-shot selection methods - random sampling, semantic embedding, and TF-IDF vectors - and evaluate these methods across multiple LLMs, including GPT-4o, GPT-3.5-turbo, DeepSeek-V3, Gemma-3, LLaMA-3…

Agentic UAVs: LLM-Driven Autonomy with Integrated Tool-Calling and Cognitive Reasoning
Anis Koubaa, Khaled Gabr
https://arxiv.org/abs/2509.13352 https://ar…

Agentic UAVs: LLM-Driven Autonomy with Integrated Tool-Calling and Cognitive Reasoning
Unmanned Aerial Vehicles (UAVs) are increasingly deployed in defense, surveillance, and disaster response, yet most systems remain confined to SAE Level 2--3 autonomy. Their reliance on rule-based control and narrow AI restricts adaptability in dynamic, uncertain missions. Existing UAV frameworks lack context-aware reasoning, autonomous decision-making, and ecosystem-level integration; critically, none leverage Large Language Model (LLM) agents with tool-calling for real-time knowledge access. …

HumorPlanSearch: Structured Planning and HuCoT for Contextual AI Humor
Shivam Dubey
https://arxiv.org/abs/2508.11429 https://arxiv.org/pdf/2508.11429

HumorPlanSearch: Structured Planning and HuCoT for Contextual AI Humor
Automated humor generation with Large Language Models (LLMs) often yields jokes that feel generic, repetitive, or tone-deaf because humor is deeply situated and hinges on the listener's cultural background, mindset, and immediate context. We introduce HumorPlanSearch, a modular pipeline that explicitly models context through: (1) Plan-Search for diverse, topic-tailored strategies; (2) Humor Chain-of-Thought (HuCoT) templates capturing cultural and stylistic reasoning; (3) a Knowledge Graph to r…

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[3/10]:
- Context-aware Biases for Length Extrapolation
Ali Veisi, Hamidreza Amirzadeh, Amir Mansourian

Replaced article(s) found for cs.CL. https://arxiv.org/list/cs.CL/new
[3/9]:
- SoLoPO: Unlocking Long-Context Capabilities in LLMs via Short-to-Long Preference Optimization
Sun, Liao, Han, Bai, Gao, Fu, Shen, Wan, Yan, Zhang, Huang

Do LLMs "Feel"? Emotion Circuits Discovery and Control
Chenxi Wang, Yixuan Zhang, Ruiji Yu, Yufei Zheng, Lang Gao, Zirui Song, Zixiang Xu, Gus Xia, Huishuai Zhang, Dongyan Zhao, Xiuying Chen
https://arxiv.org/abs/2510.11328

Do LLMs "Feel"? Emotion Circuits Discovery and Control
As the demand for emotional intelligence in large language models (LLMs) grows, a key challenge lies in understanding the internal mechanisms that give rise to emotional expression and in controlling emotions in generated text. This study addresses three core questions: (1) Do LLMs contain context-agnostic mechanisms shaping emotional expression? (2) What form do these mechanisms take? (3) Can they be harnessed for universal emotion control? We first construct a controlled dataset, SEV (Scenari…