arXiv:2608.20794v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) can degrade factual behavior outside the target domain. This degradation is…
Tag: cs.AI updates on arXiv.org
Structure for Reading, Prose for Writing: Asymmetric Structural Conditioning in Multi-Agent Document Authoring
arXiv:2608.20786v1 Announce Type: new Abstract: Multi-agent pipelines that author formal documents must both read a requester’s forms and write against…
Beyond Endpoint Gains: A Weight-Delta Audit of Medical Specialization
arXiv:2608.20768v1 Announce Type: new Abstract: Specialist language models are usually understood through endpoint gains: the generalist scores lower, the…
Continuous-Time Quantum Walks based Graph Neural Network
arXiv:2608.20738v1 Announce Type: new Abstract: Graph Neural Networks (GNNs) are widely used on graph-structured data, but most suffer from two key…
Is Multimodal Speculative Decoding Ready for Diffusion-Based Parallel Drafting? A Survey and Empirical Diagnosis
arXiv:2608.20743v1 Announce Type: new Abstract: Speculative decoding accelerates autoregressive generation by allowing a lightweight drafter to propose…
ForeTime-VLA: Causal Future-Token Distillation from a World Action Model for Conveyor-Belt Manipulation
arXiv:2608.20735v1 Announce Type: new Abstract: Manipulating moving objects requires a policy to anticipate contact events, yet vision-language-action…
Natural-Language-Guided Generator-Agnostic Shortlisting for Protein Binder Design
arXiv:2608.20755v1 Announce Type: new Abstract: Modern de novo design workflows generate many candidate protein binders, but wet-lab validation capacity…
Calibrating Criterion Revision in LLM Agents: Failure Modes and a Trace-Anchored Protocol
arXiv:2608.20729v1 Announce Type: new Abstract: Language-model agents can improve after failure or carry text across episodes without revising what counts…
VortexChat: An agentic framework for autonomous multi-objective integrated photonic design
arXiv:2608.20688v1 Announce Type: new Abstract: The advancement of modern integrated photonics is frequently bottlenecked by device design workflows that…
DreamBench-SWE: A Multi-Session Memory-Hygiene Benchmark for Software Agents
arXiv:2608.20664v1 Announce Type: new Abstract: DreamBench-SWE is a multi-session benchmark for software-agent memory hygiene in which later software…