Evolve Quick Reviews
Public Paper Reviews
Structured preliminary reviews of recent open-access papers, generated and continuously updated by the Evolve analysis engine. Each review condenses what the paper claims, how solid the evidence is, and the questions a rigorous peer reviewer would ask — English analysis with abstracts in your language. These are independent AI analyses, not peer review; the canonical record always lives at the publisher.
Vox-Infinity: Benchmarking the Limits of Long-Context Spoken Language Models
Xize Cheng, Wenxu Jia, Chenyuhao Wen, Dongjie Fu et al.
Claims first provenance-annotated long-context speech benchmark; recency effect solid, but LLM-built QA, LLM judging and partial data release limit rigor.
CodeMidas: Scaling Agentic Coding RL Environments from Code Itself
Bowen Ye, Lei Li, Shicheng Li, Zihao Yue et al.
Claims source code alone can yield 5,545 RL environments that lift MiMo-V2.5 on five benchmarks; gains plausible but single-run, no significance tests.
Structuring occupational accident narratives for cross-sector safety analysis: Transferability of accident-process role classification
Aho Yapi, Pierre Latouche, Arnaud Guillin, Yan Bailly
Claims unit-level accident-role classifiers transfer across sectors (~85.7% OOD balanced accuracy); rigorous, honest, but data/code unavailable.
Designer-RSI: Evolving Procedural Memory from User Traffic for Agentic Graphic Design
Hongyang Du, Lan Yan, Christian Flores, Asim Kadav
Skill bank evolved from user traffic lifts a frozen design agent (GenEval2 72.7→99.3%, Sonnet-4); credible but one-judge evaluation, no method baselines.
Connected Content Retriever: Dense Graph Edge Features Powering Pre-Ranking at LinkedIn
Akhilesh Gupta, Sudarshan Srinivasa Ramanujam, Chirag Bhanuprasad Mehta, Reshma Asharaf Beena et al.
LinkedIn's GPU pre-ranker with runtime edge-feature joins lifts Feed time spent +2.5%; solid systems work, but training and baseline details are thin.
Apollo Restore: A Foundation LLM for Historical Greek Optimized for Fill-in-the-Middle Restoration of Ancient Greek Texts
Hope McGovern, Anna Dolganov, Samuel Belkadi, Guillaume Kunsch et al.
First 24B FIM model for Ancient Greek gap restoration; SOTA claims over Ithaca/Cullhed look credible, but no significance tests and some figures provisional.
Toward Personalized Sleep Guidance from Wearable Data Using Language Models
Yusheng Tan, Running Zhao, Sofia Angel, Ninghui Hao et al.
Distills multi-agent sleep guidance into 0.8B SLMs that beat commercial LLMs on self-referenced metrics; evaluation circularity tempers the headline claims.
Efficient Mixture-of-Experts with Speculative Decoding via Expert Coactivation
Kumari Nishu, Han-Byul Kim, Santosh Chilkunda, Maxwell Horton et al.
Apple trains MoE routers for expert coactivation, claiming 21% SD throughput gain at on-par accuracy; solid ablations but single-seed, ~1B-scale evidence only.
Goal-driven Variant Categorization
Daniel Calegari, Daniel Amyot
Reverses variant clustering via a declared GRL goal-model axis; LLM categorization is novel, transparently evaluated, yet confounded and descriptive only.
A Governance-Aware Large Language Model Orchestrated Agentic Digital Twin for Transmission System Operator Control Room Decision Support
Costas Mylonas, Magda Foti, Emmanouel Varvarigos
Claims four runtime safety invariants for LLM grid-control agents hold on all 2006 benchmark runs; rigorous audit-trail evaluation, but only a 35-bus network.
Machine Learning for Invisible Dark Boson Searches at the Electron-Ion Collider
Rojae Mighty, Ankush Reddy Kanuganti
Rigorous generator-level comparison: BDTs match optimized cuts with electron-only inputs; adding t gives a small but replica-robust BDT gain at 10 GeV.
Replication Without Persistence in Hosted LLMs: Measurement Sensitivity in Action-Time Belief Evaluation
Bhushan Kashinath Joshi
复现成功而持续性失败：三问分离论证严谨，但单环境、捆绑混淆与依赖支持度的指标限制了结论范围
Semantics Delivery Network: Rethinking Web Retrieval Infrastructure for LLM Agents
Peichun Hua, Yunming Xiao
Proposes SemDN, a CDN-like edge network delivering chunks to LLM agents; motivating measurements are solid but the core coverage mechanism is untested.
CultureMINE: Datasets and Methods for Improving the Cultural Capabilities of NLP Systems
Tania Chakraborty, Eylon Caplan, Zhaoqing Wu, Kevin Cushing et al.
Survey of >375 cultural NLP papers gives a 9-capability taxonomy; data work outpaces systems work. Useful resource, but annotation reliability is unquantified.
COREM: Cosine-Relation Momentum Reshaping with Stateful Writeback
Yan Wang, Xiaochuan Wang, Yuxiang Sun
COREM reshapes momentum via cosine relations; competitive with Muon at toy scale (better CIFAR-10, par on enwik8) but scope and ablations limit claims.
EmbeddGAN: A Novel GAN Framework Using an Embedding Network and Gini Distance Correlation
MaTais Caldwell, Yixin Chen, Xin Dang, Charles Walter
EmbeddGAN swaps the GAN discriminator for a gCor-maximizing embedding net; competitive but never best FID, single-run evidence, code unmentioned.
Defusing Explosive Prompts: Understanding and Preventing Trigger-Based Prompt Injections in LLM Agents
Justin Szczepaniak, Elad Feldman, Naum Viner, Ben Nassi
Dormant trigger-based prompt injections beat IPI defenses across models and agents; evidence is broad and candid but in-distribution and two-turn.
The Wisdom of Artificial Deliberative Crowds
Federico Barrera-Lemarchand, Mariano Sigman, Joaquin Navajas
Across 4 domains deliberating LLM crowds beat averaging only with model diversity; paired evidence strong but bootstrap p-values inflated, code unreleased.
When Cosine Similarity Fails to Reflect Linearly Accessible Structure in Dialogue Models
Yu Sun, Mengyin Lu, Cong Feng, Guangming Lu et al.
Argues cosine similarity underestimates linearly decodable persona structure in dialogue LLMs; careful, well-controlled empirical study with modest novelty.
Agreement Overstates Evidence: Error Dependence in LLM Judge Consensus
Elias Hossain, Niloofar Yousefi, Ser-Nam Lim
Claim: correlated errors make LLM-judge consensus overstate evidence (10 judges ≈ 3.5 independent). Well-supported, candid about limits, though headline ρ partly reflects shared item difficulty.
FRAMES: Failure Recovery And Monitoring of Embodied Skills for Humanoid Loco-Manipulation
Ajay Vikram Periasami, Xinyuan Luo, Haoyu Li, Xianyi Cheng
FRAMES adds an LLM/VLM supervisory layer for humanoid loco-manipulation; only the monitor is evaluated (94% accuracy, 100 sim trials); recovery stays untested.
IntLawNER: A Named Entity Recognition Dataset and Benchmark in International Law
Genis Skura, Roland Bouffanais, Didier Wernli
First international-law NER benchmark (2,987 sentences, 7 types); sound pipeline and error analysis, but Claude-silver circularity and rare-label sparsity.
Strategic Classification Has a Missing Lever: Audit Risk
Raman Ebrahimi, Massimo Franceschetti
Extends strategic classification with per-feature audits; decomposition, DkS-hardness, regret rates are rigorous; commitment assumption narrows scope.
EvidenT: Building Trustworthy Enterprise Assistants through Evidence Groundedness and Traceability
Anubha Kabra, Katie Jooyoung Kim, Colin Zhiwei Kou, Helene Sajer et al.
Training-free n-gram verification lifts enterprise RAG groundedness (77-93% vs 47-73%) but traceability wins are partly by construction.