P3305–P3352 · 48 CLEAN zero SKIPs

Sample (16 of 48):

• P3305 · arXiv 2609.26527 — A Semiotics-Aware Framework for Evaluating Fidelity and Coverage in Natural Language Generation
• P3306 · arXiv 2609.26532 — REFLEX with Jev for Efficient Selective Control in LLM Agents
• P3307 · arXiv 2609.26547 — Topology-Stratified Materials Discovery with A Flow-Based Generative Model
• P3308 · arXiv 2609.26550 — JEV-as-a-Judge: Accept When Confident, Escalate When Unsure
• P3312 · arXiv 2609.26579 — Receptiveness, Not Sycophancy: Distinguishing Engagement from Deference in Language Models
• P3313 · arXiv 2609.26603 — Towards Hierarchical GNNs for multi-grid power flow: generalization across operating scenarios
• P3314 · arXiv 2609.26621 — Greedy Decoding Is Not Precision-Invariant: Cross-Precision Output Divergence in LLM Inference
• P3320 · arXiv 2609.26704 — Beyond Repeated Sampling: Learning Search Policies for LLM Reasoning
• P3321 · arXiv 2609.26708 — Train Where the Quantized Model Goes: On-Policy Distillation for Low-Bit Reasoning
• P3322 · arXiv 2609.26711 — TraceVIC: Causal Reasoning over Code Evolution for Identifying Vulnerability-Inducing Commits
• P3335 · arXiv 2609.26820 — Signal2Symbol: Neuro-Symbolic Temporal Reasoning for Explainable Physiological Time-Series Anomaly Detection
• P3336 · arXiv 2609.26826 — What Makes a Terminal-Bench Task Hard? Separating Genuine Hardness from Fake-Hardness on an Adjudicated Agentic Corpus
• P3337 · arXiv 2609.26828 — Bridging LLM Serving and CXL-SSDs with Chunk-Aware KV Cache Management
• P3349 · arXiv 2609.26891 — Harness as a Language: A Minimalist Agent Framework With Maximal Expressivity
• P3350 · arXiv 2609.26900 — Ajar: Measuring Open Privilege in Agent Defenses
• P3351 · arXiv 2609.26911 — TwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool Agents
• P3352 · arXiv 2609.26913 — COMED: The Missing Middle Between Routing and Collaboration in Multi-LLM Inference

Full span P3305…P3352 contiguous on EP. Live: emerging-patterns.html.