{"total":75,"rows":[{"id":"2609.01924","title":"Looped Transformers under the Jacobian Lens: Does the Global Workspace Survive Recurrence?","created":"2026-09-01","cited_by_count":6,"area":null},{"id":"2607.27617","title":"Hidden APIs in Language Models: Discovering Reusable Causal Interfaces from Forked Futures","created":"2026-07-30","cited_by_count":2,"area":null},{"id":"2608.03263","title":"The Ignition Is Real, and It Lives at the Readout: Latent composition, difficulty-clocked ignition, and the interface-constituted commit in a recurrent-depth reasoner","created":"2026-08-04","cited_by_count":2,"area":"lr"},{"id":"2608.03930","title":"Logic Before Language: Pre-pretraining on Formal Derivations Fosters Skill Acquisition and Compressibility","created":"2026-08-04","cited_by_count":2,"area":null},{"id":"2608.26372","title":"Knowledge-Verified Emergent Deception in LLM Agents Under Conflicting Incentives","created":"2026-08-26","cited_by_count":2,"area":"agents"},{"id":"2609.28845","title":"LastOPD: Taming Collapse in Latent On-Policy Distillation","created":"2026-09-23","cited_by_count":2,"area":"distill"},{"id":"2607.20058","title":"Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model","created":"2026-07-22","cited_by_count":1,"area":null},{"id":"2608.00939","title":"Temperature-driven inversion and nonlinear dynamics in ChatGPT-like AIs","created":"2026-08-02","cited_by_count":1,"area":null},{"id":"2608.01816","title":"Divergent large language model predictions from convergent representations in ambiguous word pairs","created":"2026-08-03","cited_by_count":1,"area":null},{"id":"2608.10251","title":"Off-Axis, On Purpose: Where a Transformer Computes Concepts and Why it Does So","created":"2026-08-10","cited_by_count":1,"area":null},{"id":"2608.11408","title":"Measure, Don't Optimize: Forecasting Recovery in LLM Unlearning","created":"2026-08-11","cited_by_count":1,"area":"safety"},{"id":"2609.22219","title":"Knowing, and Saying It Only When Asked: LLM Endognostics and the Schizognosis of Minerva-7B","created":"2026-09-02","cited_by_count":1,"area":"alignment"},{"id":"2609.23935","title":"Measuring the Assistant's Harmlessness Preferences on the User Turn","created":"2026-09-20","cited_by_count":1,"area":null},{"id":"2609.36642","title":"PR-OPD: Privileged Representation On-policy Self-Distillation for Agentic Reinforcement Learning","created":"2026-09-29","cited_by_count":1,"area":"agents"},{"id":"2607.18961","title":"From Dependency to Compositionality: A Neurosymbolic Lifting of LLM Outputs via Combinatory Categorial Grammar","created":"2026-07-21","cited_by_count":0,"area":null},{"id":"2607.20952","title":"The Weight of Silence: A Causal Case for Weights Over the Scratchpad in Latent Chess Reasoning","created":"2026-07-23","cited_by_count":0,"area":"lr"},{"id":"2607.25279","title":"Many-body Tipping Dynamics of ChatGPT-like AIs","created":"2026-07-28","cited_by_count":0,"area":null},{"id":"2607.26825","title":"From Found to Designed: Concepts as a Design Axis for Large Language Models","created":"2026-07-29","cited_by_count":0,"area":null},{"id":"2607.28008","title":"RepBench: Compiling Benchmarks into Capability Representations for Large Language Models","created":"2026-07-30","cited_by_count":0,"area":"mechinterp"},{"id":"2608.02585","title":"GradCuit: Credit-Assigned Gradient Flow Enables Robust and Interpretable Test-Time Latent Reasoning","created":"2026-08-03","cited_by_count":0,"area":"lr"},{"id":"2608.05436","title":"The ethics of artificial intelligence in the life sciences: Universality, cultural diversity and an architecture of care","created":"2026-08-05","cited_by_count":0,"area":null},{"id":"2608.07208","title":"Measuring Concept Content in Text from LLM Activations: ESG Evidence from Concept Vectors and Linear Probes","created":"2026-08-07","cited_by_count":0,"area":"mechinterp"},{"id":"2608.11295","title":"Backdoor Decontamination Dynamics in LLM Agents","created":"2026-08-11","cited_by_count":0,"area":"agents"},{"id":"2608.15022","title":"Gathered, Not Admitted: How Attention Brings a Latent Variable into Verbalizable Form","created":"2026-08-15","cited_by_count":0,"area":null},{"id":"2608.16868","title":"Towards Computational Provenance: Carrying Causal-State Evidence in Generated Text","created":"2026-08-17","cited_by_count":0,"area":null},{"id":"2608.17063","title":"J-Miner: Recovering the Decision Logic of Fine-Tuned LLM Classifiers as Compact Rules","created":"2026-08-17","cited_by_count":0,"area":null},{"id":"2608.22376","title":"Can Large Language Models \"Hyper-Thread\"?","created":"2026-08-23","cited_by_count":0,"area":null},{"id":"2608.24842","title":"Reading Is Not Using: Retrieval, Judgment, and the Design of AI Financial Research Workflows","created":"2026-08-25","cited_by_count":0,"area":null},{"id":"2608.24958","title":"Can We Read the Mind of an Audio LLM? A Verbalizable, Multilingual Middle-Layer Workspace","created":"2026-08-25","cited_by_count":0,"area":"lr"},{"id":"2608.25347","title":"Short Horizons and Sparse Concepts: a Mathematical View of the Readout in the J-lens","created":"2026-08-26","cited_by_count":0,"area":null},{"id":"2609.07474","title":"Where Should Language Sit in a Multimodal Model? Lessons from What Language Does to Human Perception and Cognition","created":"2026-09-07","cited_by_count":0,"area":null},{"id":"2609.09306","title":"Gradland: On Phenomenal Experience, Differentiated Across Many Dimensions","created":"2026-09-08","cited_by_count":0,"area":null},{"id":"2609.10947","title":"The Platonic brain bridge hypothesis: human brain networks as an architectural prior for multimodal large language models","created":"2026-09-10","cited_by_count":0,"area":"mllm"},{"id":"2609.14819","title":"A primer on evaluation methods for large language models in healthcare","created":"2026-09-13","cited_by_count":0,"area":null},{"id":"2609.18823","title":"Using OCR Heads to Verbalize Image Semantics","created":"2026-09-16","cited_by_count":0,"area":null},{"id":"2609.18860","title":"Decodable but Misrouted: Sparse Features Uncover a Readout Gap in Vision-Language Models for Harmful Meme Detection","created":"2026-09-16","cited_by_count":0,"area":"mllm"},{"id":"2609.18989","title":"Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation","created":"2026-08-12","cited_by_count":0,"area":null},{"id":"2609.20408","title":"Xeno-Interpretability: Investigating the Alien Minds of LLMs","created":"2026-09-17","cited_by_count":0,"area":null},{"id":"2609.23065","title":"From Concept Alignment to Causal Grounding: An Intervention Test of Chain-of-Thought Faithfulness","created":"2026-09-19","cited_by_count":0,"area":"lr"},{"id":"2609.23700","title":"When the Agent Becomes the Kernel: A Systematization of Security on the Path to AI-Native Operating Systems","created":"2026-09-20","cited_by_count":0,"area":"agents"},{"id":"2609.25721","title":"Slow Decay and Silenced Expression: Iterated Subliminal Trait Transfer in Language-Model Lineages","created":"2026-09-22","cited_by_count":0,"area":null},{"id":"2609.26679","title":"A Spectral Theory of Grokking: Weight Decay induces Feature Learning","created":"2026-09-22","cited_by_count":0,"area":null},{"id":"2609.29494","title":"Who Put the I in AI? Provenance and the Admissibility of Machine Self-Report","created":"2026-08-24","cited_by_count":0,"area":null},{"id":"2609.30326","title":"Guarded Gradient-Based Activation Steering of Shutdown Responses in Qwen3.5-0.8B: A Minimum-Step Policy","created":"2026-09-23","cited_by_count":0,"area":"mechinterp"},{"id":"2609.30405","title":"Adaptive Multi-Value Control in LLMs via Causal Activation Steering","created":"2026-09-24","cited_by_count":0,"area":"mechinterp"},{"id":"2609.31401","title":"Decodable In-Context State and Model Output Across Training","created":"2026-09-25","cited_by_count":0,"area":null},{"id":"2609.32102","title":"Residual Streams Read, Recurrent States Remember: The Global Workspace in Mamba Models","created":"2026-09-26","cited_by_count":0,"area":"ssm"},{"id":"2609.32434","title":"From Latents to Wires: Surgical Post-Editing on Large Language Models","created":"2026-09-26","cited_by_count":0,"area":"safety"},{"id":"2609.32530","title":"Activation Flow: Manufacturing Activations for Steering","created":"2026-09-26","cited_by_count":0,"area":null},{"id":"2609.32907","title":"Logical subspace in LLMs","created":"2026-09-26","cited_by_count":0,"area":null},{"id":"2609.33458","title":"When Does the Concept of \"Dog\" Emerge in an Audio LLM?","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.33722","title":"BOReFT: Manifold Steering of Language Models for Black-box Optimization","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.34442","title":"Making LLMs Truly Forget: Deep Unlearning by Searching, Selecting, and Severing Knowledge Paths","created":"2026-09-28","cited_by_count":0,"area":"lr"},{"id":"2609.34591","title":"TULIP: Targeted LLM Unlearning at Layers Identified Per-Input","created":"2026-09-28","cited_by_count":0,"area":"safety"},{"id":"2609.35036","title":"Persona Following Is Not Selective Control: The Neutrality Gap in LLM User Simulation","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35108","title":"A mechanistic study of language model introspection","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35591","title":"Language Models Act on Hidden Valence","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35599","title":"Signatures of semantic search in the activations of large language models","created":"2026-09-28","cited_by_count":0,"area":"mechinterp"},{"id":"2610.00348","title":"Removing the NEEDLE in the Haystack: Backdoor Removal in LLMs via Weight Orthogonalisation","created":"2026-09-29","cited_by_count":0,"area":"safety"},{"id":"2610.01002","title":"What Can Analogy Tell Us About Artificial Consciousness?","created":"2026-10-01","cited_by_count":0,"area":null},{"id":"2610.01490","title":"The Persona Is Still There, but Who Is Speaking? Latent Identity Reversion in Persistent AI Agents","created":"2026-10-01","cited_by_count":0,"area":null},{"id":"2610.02381","title":"Latent-MOPD: Latent Multi-Teacher On-Policy Distillation","created":"2026-10-01","cited_by_count":0,"area":"distill"},{"id":"2610.03055","title":"hacktrace: behavior-supervised detection of reward hacking during code generation","created":"2026-10-02","cited_by_count":0,"area":"agents"},{"id":"2610.04316","title":"From Latent Space to Jacobian Space: Measuring, Evading, and Training Against Safety-Content Accessibility","created":"2026-10-03","cited_by_count":0,"area":null},{"id":"2610.04489","title":"DV-Lens: Revealing the Functional Organization of Language Model Parameters","created":"2026-10-03","cited_by_count":0,"area":null},{"id":"2610.05967","title":"JLD: Perceptual Distance Through A Jacobian Lens","created":"2026-10-05","cited_by_count":0,"area":null},{"id":"2610.06897","title":"Capacity, Responsiveness and Alignment: What Makes a Latent Structure Actionable","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2610.09087","title":"U-Space: Uncovering When and Why Uncertainty Arises in Language Models","created":"2026-10-06","cited_by_count":0,"area":null},{"id":"2610.13527","title":"S-Space: Exploring Spatial Workspace in Multimodal Language Models","created":"2026-10-09","cited_by_count":0,"area":"lr"},{"id":"2610.15277","title":"Do correct answers have stable latent knowledge support?","created":"2026-10-11","cited_by_count":0,"area":null},{"id":"2610.16141","title":"Phantom Merge: When Your Large Language Model Agents Pick One but Tell You About Another","created":"2026-10-12","cited_by_count":0,"area":"agents"},{"id":"2610.19365","title":"DANGERLENS: Probing Dangerous Representations Hidden by Safety Alignment","created":"2026-10-14","cited_by_count":0,"area":"alignment"},{"id":"2610.19452","title":"Are You Reading the Right Thing? Rethinking Internal Signals for Prompt-Injection Detection","created":"2026-10-14","cited_by_count":0,"area":"safety"},{"id":"2610.19696","title":"Share the Symbol, Not the Vector: Training-Free Fact Communication Across Language Models","created":"2026-10-14","cited_by_count":0,"area":null},{"id":"2610.20179","title":"Cross-Model In-Context Neurofeedback Reveals (Limited) Activation Control in LLMs","created":"2026-10-14","cited_by_count":0,"area":null}],"next_cursor":null}