Every paper checked — nothing hidden

All papers

Every paper checked in the week of 12–18 Oct 2026 · LLMs for code — nothing is hidden. Brief papers first.

All1,106 In brief3 More (on-topic)144 Not relevant928 Not assessed6 Older42 Could not confirm7 Blogs with no paper39 All blogs42

Paper	Topic	Relevance
Soliduzz: Finding Miscompilations in Smart-Contract Compilers with Assertion-Driven Equivalence Oracles Finds miscompilations in smart-contract compilers using executable assertions and standard fuzzers. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on compiler fuzzing rather than LLM code assistants.
sGLP: Specifying Stochastic Agents that Simulate People Presents a stochastic logic programming method to simulate human user behavior for testing decentralized grassroots apps. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Presents a stochastic simulation language, not LLMs writing code.
Grounding Deep Learning Framework Testing in Data Flow This paper introduces a data-flow graph representation and mutation testing technique to uncover bugs in deep learning frameworks. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Tests deep learning execution frameworks, not code generation or assistants.
HybridDualPathNet: Cross-Fusion Attention of EfficientNetV2 and Swin Transformer for Brain MRI Classification of Alzheimer's and Parkinson's Disease The authors introduce a hybrid architecture combining EfficientNetV2 and Swin Transformer to classify Alzheimer's and Parkinson's disease from brain MRI. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Medical image classification; cs.SE category appears misapplied.
DICE: Tilting the Odds through Interface Attackson System One Model This paper analyzes vulnerabilities in system interfaces where adapter substitutions alter model decisions without modifying the underlying model itself. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes pipeline interface attacks, not code generation or coding assistants.
(Implicitly) Characterizing Probabilistic Polynomial Time as Used in Cryptography The paper formalizes probabilistic polynomial-time cryptographic computation and adversary reductions using an implicit lambda calculus framework. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Theoretical lambda-calculus for cryptography, not software engineering with LLMs.
To Cache or Not to Cache: Understanding and Mitigating Safety Incompatibility in Cross-Model KV Cache Reuse The paper studies safety behavioral deviations caused by cross-model key-value cache reuse and proposes an alignment technique to mitigate them. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Investigates KV cache reuse across models, not software engineering.
X.js: A Live Benchmark for Node.js Taint-Style Vulnerabilities The paper presents X.js, a benchmark of real-world Node.js taint-style vulnerabilities paired with executable exploits and patches. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Introduces a Node.js security vulnerability dataset, not LLMs for code or software agents.
Fin_Agent at FinEval 2026: Task-Aware Methods for Multilingual Financial NLP The paper describes a system submission combining rules and language models for multilingual financial named-entity recognition and question answering. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on financial text tasks like NER and QA, not software engineering or code.
Does System One Transfer? Calibrated Typed-Decision Models, Label Budgets, and Translation for Indonesian Regional Languages The paper evaluates the cross-lingual transfer, calibration, and translation sensitivity of fast typed-decision models on regional Indonesian languages. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Evaluates regional language sentiment classification, not code generation or repair.
The AI's Poker Face: Contrasting Persuasion and Sycophancy in Voice-Based Multi-Turn Conversations The paper studies opinion change, rhetorical strategies, and sycophancy in spoken multi-turn dialogues between humans and language models. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes spoken dialogue on social topics, not code generation or software engineering.
OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools The paper presents OpenMobile-2, combining open mobile environments, trajectory datasets, and hybrid GUI-tool actions for mobile agents. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on general mobile application workflows rather than software engineering or code.
Simple Linear Corrections for Compression in Late Interaction The paper proposes a linear score correction method to recover retrieval quality in compressed late-interaction document representations. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on information retrieval scoring, not programming or code generation.
Modeling the Arrival Time of Information-Bearing Events in Natural Language The paper uses neural temporal point processes to model the timing of information-bearing events in human and model generated text. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Investigates natural language generation properties, not software engineering.
The Thinking Leakage in Reinforcement Learning for LLMs The paper examines whether non-thinking mode reinforcement learning gains in hybrid reasoning models stem from leakage of thinking-mode activations. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes general reasoning models rather than programming tools.
ASSET: Adaptive Speech Segmentation via Reinforcement Learning The paper introduces ASSET, a reinforcement learning method that adaptively compresses speech encoder outputs into continuous audio tokens. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses speech encoding, not software code generation.
Evaluating LLMs as Agentic Workflow Executors in Real-World Customer Support Skills The paper introduces a benchmark derived from real customer support dialogues to evaluate language models executing bounded, tool-using skills. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Evaluates customer support dialogues, not programming assistants.
Multimodal Co-Generation With Continuous Bitstream Diffusion The paper applies continuous bitstream diffusion to joint and conditional generation across multimodal data without learned tokenizers. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on multimodal generative diffusion, not software coding.
Tokenization-Invariant Watermarking: A Byte-Local Method for LLM Text The paper presents a byte-local text watermarking method that scores outputs without invoking tokenizers during detection. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses text provenance watermarking, not code generation.
SciCurateBench: Curating Experimental Data from Heterogeneous Scientific Literature with Verbal-Action Agents The paper presents SciCurateBench to evaluate language models on verifying complex experimental conditions from unprocessed scientific literature. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on scientific literature curation rather than software engineering.
Scaling Laws for Multimodal Data Mixtures The paper derives multimodal data mixing scaling laws across text, vision, and audio using mixture-of-experts models. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on multimodal data scaling, not code generation.
MemForm: Benchmarking Memory Formation Poisoning in LLM Agents The paper introduces MemForm to benchmark adversarial memory formation poisoning across five input interfaces in language model agents. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies agent memory vulnerabilities, not software coding.
Neither True nor False: Do LLMs Represent Truth-Value Gaps in Plural Predication? The paper introduces HomogeneityNLI to evaluate whether open-weight language models recognise truth-value gaps in plural predication. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Examines natural language semantics, not software engineering models.
Evaluating Semantic Relation Knowledge through Bidirectional Relation Prediction from Gloss Pairs The paper introduces a bidirectional gloss-pair prediction task to evaluate semantic relation knowledge across multiple languages in language models. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Tests lexical semantic consistency, not programming assistants.
Evaluating Bias in Chinese Large Language Models across Tasks and Registers: Task Operationalization Effects and Irony Classification Risks This paper audits Chinese language models for bias and error propagation across formal and ironic social registers. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies content moderation tasks, not code generation or repair.
EdgeLLM-Nano: Auditing Polar-Guided Key-Cache Compression and Consistency-Anchored Pruning on a 33.6 MB On-Device Language Model This paper audits key-cache compression and weight pruning techniques on a small edge-deployed language model testbed. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes on-device model compression, not software coding tools.
Query-Adaptive Neuron Steering for LLMs This paper proposes a framework that dynamically steers internal language model neurons based on query-induced activation patterns. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on general inference steering rather than coding tasks.
LongHisDoc: A Comprehensive Benchmark for Chinese Long Historical Document Understanding This paper introduces a benchmark of expert-annotated question-answer pairs to evaluate multimodal comprehension of long historical Chinese documents. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on historical document comprehension, not programming tasks.
Reward-Aligned Distillation This paper presents a bilevel distillation method that weights teacher supervision tokens based on their predicted impact on student reward. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Evaluates math reasoning benchmarks, not software engineering or code.
Uniform Information Density-Based Preference Optimization This paper proposes a preference optimization algorithm grounded in uniform information density theory to align language model outputs. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses general alignment, not coding assistants or software repair.
Fine-Tuning Strategy, Not Just Data: A Statistically Rigorous Audit of Gender and Stereotypical Bias in Embedding Models for HR Recommendation Systems This paper audits gender and stereotypical bias in multilingual job-matching embedding models using robust statistical tests. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies HR embeddings, not code generation or software engineering.
The Company a Question Keeps: Auditing Open Jev-Style Models for Retrieval This paper audits request independence in open Jev-style retrieval models across multi-passage query states. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies information retrieval models, not software development.
PVFH-Bench: Benchmarking Premise-Induced Visual Fixation Hallucinations in Vision-Language Models This paper introduces a benchmark to measure vision-language model hallucinations caused by deceptive textual premises paired with visually confusable images. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies vision-language hallucinations, not software engineering.
VoTSpeech: Decoupling Voice Design from Speech Generation in Instruction-Following TTS This paper presents an instruction-following text-to-speech model that separates voice design from speech generation using an intermediate voice latent. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Deals with acoustic modeling, not software development.
Amharic-English Cross-Language Question Answering Using Statistical Machine Translation, Named Entity Recognition, and Semantic Indexing This paper presents a bidirectional Amharic-English cross-language question answering system using machine translation, entity recognition, and semantic indexing. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses cross-lingual QA, not software engineering.
The Constitution Speaks Nheengatu: An Open MT System and Reproducible Corpus Pipeline for the Amazonian Língua Geral This paper releases an open parallel corpus, speech evaluation set, and translation models for Portuguese and the Amazonian language Nheengatu. Peer-reviewed	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses low-resource natural language translation, not code.
Who Receives Respect? A Counterfactual Benchmark for Honorific Allocation Bias in Bengali This paper introduces a counterfactual benchmark to evaluate honorific register allocation bias across gender and occupations in English-to-Bengali translation. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Examines machine translation bias, not software engineering.
Training-Free BioNEL-RAG: Advancing Biomedical Entity Linking through Retrieval-Augmented Generation and Large Language Model This paper introduces a training-free retrieval-augmented generation framework for biomedical named entity linking to improve cross-domain adaptability. Peer-reviewed	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses biomedical entity linking, not code generation.
Dependency Reasoning for Aspect Sentiment Quadruple Prediction This paper injects linearized dependency parse trees into language model prompts to improve aspect sentiment quadruple prediction. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on sentiment quadruple extraction, not software code.
Social Bias in Chinese Large Language Models under Social Media Expressions This paper evaluates social bias in Chinese language models across social media expressions, internet slang, and sarcastic packaging. Peer-reviewed	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Investigates social bias, not software code generation.
Agent-Guided Evidence Acquisition for Emotion-Cause Pair Extraction This paper investigates an agent framework that selectively queries expert judgments for borderline emotion-cause pairs to improve extraction accuracy. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses emotion-cause pair extraction, not software development.
Description Complexity Bounds for Analytic Approximations of Lipschitz Functions This paper derives theoretical description complexity bounds for approximating one-dimensional Lipschitz functions using elementary analytic expression trees. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Mathematical approximation theory, not software engineering.
TeachGrade: Training-Free Short Answer Scoring with Teacher Evidence This paper proposes a training-free framework that uses teacher-graded examples and dual memories to improve automated short answer grading. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Assesses short answer scoring, not software engineering.
FidelityQ: Preserving Predictive Fidelity in Ultra-Low-Bit Language Model Quantization This paper introduces a scalar quantization method using analytic coding to preserve language model predictive fidelity at precisions down to two bits. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on weight quantization, not coding tasks.
Do Personalized LLMs Need the Whole Conversation? Matching the Full History with Only 5% of Its Text via Relevance-Filtered Context This paper demonstrates that sentence-level relevance retrieval can match full conversation history accuracy in personalized language models using only five percent context. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Examines conversation history retrieval, not code generation.
HiPAD: A Unified Gait, Audio-Visual, and Personality Fusion Network for Depression Detection, with a Hierarchical Personality Ablation and the Calibration Limits of Domain-Adapted LLMs This paper presents a multimodal fusion network for depression detection combining gait, audio-visual data, and textual personality descriptions. Peer-reviewed	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on depression detection, not software engineering.
HodgePhaseNet: Discrete Hodge Decomposition of Semantic Flows in Transformer Attention This paper introduces a discrete Hodge decomposition method to separate semantic flow transitions within transformer attention mechanisms. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on mathematical formulation of attention, not code generation.
What Drives AI Citations? Content, Retrieval and Model Knowledge in Generative Engine Optimization (GEO) This paper examines how content matching, retrieval mechanics, and model knowledge affect web page citations in generative search engines. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes generative search citations, not code generation.
PAT: A Personality-Driven Augmentation and Transfer Learning Framework for Depression Detection on Social Media This paper presents a personality-driven augmentation and contrastive learning framework to improve depression detection from social media posts. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Addresses social media mental health detection, not software engineering.
RegExtract: A Benchmark for PFAS Regulatory Information Extraction with LLMs The paper introduces an information extraction benchmark for extracting structured regulatory fields and source provenance from dense environmental policy documents. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Studies regulatory legal extraction, not software code tasks.
Lumen-bench: Cross-Lingual Safety Gaps and the Say-Do Problem in Tool-Calling Large Language Models The paper introduces a multilingual benchmark evaluating safety and discrepancies between verbal refusals and tool actions in tool-calling language models. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Evaluates tool use and refusal, not software engineering.
Compute-Optimal Scaling Laws for Spiking Language Models The paper presents an empirical scaling law study measuring parameter and compute trade-offs in spiking neural network language models. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Examines general language modeling scaling laws, not code generation.
Where Did We Go Wrong? Localizing Errors to Repair Multi-Turn Conversations The paper presents a framework to probabilistically localize user-intent misinterpretations and repair conversational states in multi-turn assistant systems. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Deals with conversational assistant state repair, not code generation.
Learning Proposition-Level Links from Sentence-Level Supervision for Neuro-Symbolic NLI The paper proposes a trainable proposition linker that learns atom-level semantic links from sentence-level supervision for neuro-symbolic inference. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Applies to natural language inference rather than software code.
Sanzaru: Benchmarking Protocol Formation under Epistemic & Communication Asymmetry The paper introduces a benchmark evaluating how multi-agent teams develop communication protocols under asymmetric information and restricted communication channels. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Focuses on cooperative puzzle communication, not software coding.
Active Learning for Finding a Needle in a Haystack The paper evaluates whether active learning strategies can overcome initial selection biases when identifying rare document classes in large text corpora. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Evaluates document classification datasets, not software engineering tasks.
Word-Aligned Prosody Reflects Speaker Differences More Than Depression or PTSD Severity The paper investigates whether word-aligned prosodic speech measurements reflect speaker identity traits more strongly than depression or PTSD severity. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Investigates speech prosody rather than software code generation.
SAFE-KV: Causal Suffix Repair for Behavior-Preserving Non-Prefix KV Cache Reuse in RAG The paper presents a repair method that re-encodes causal suffixes to enable safe non-prefix KV-cache reuse in retrieval-augmented generation. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Optimizes attention KV cache serving, not software engineering.
Do LLMs differentiate linguistic traits internally?: A layer-wise analysis The paper conducts layer-wise representational analysis to determine how language models internally distinguish conventional meanings from idiomatic traits. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Analyzes linguistic trait embeddings, not code generation.
Do correct answers have stable latent knowledge support? The paper proposes stability metrics to measure whether correct outputs in language models are backed by stable internal knowledge representations. Preprint	LLMs for code via arxiv · 15 Oct, 06:18	Not relevant Examines factual and math stability, not software code generation.

541–600 of 1,106← Previous Next →