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.
Summarization Bias: The Directional Collapse of Objective Projection into Told-Mode Labels in Large Language Models --- A Conceptual Framework and Registered Test Protocol
Levent Bulut
Pre-registered, falsifiable construct: LLMs collapse shown-mode prose into told labels. Elegant protocol design, but zero new data and only self-citations.
Relational Attention for Data-Efficient Language Modeling
Adrian Brasoveanu, Ece Takmaz, Jakub Dotla\v{c}il
BabyLM entry adds relational attention (DAT) + NextLat objective: solid BLiMP gains from architecture, NextLat effects near seed noise; controlled and honest.
Steering the Compass: Aligning Dynamic Psychological Counseling Conversations with Cognitive Behavioral Therapy Strategies
Zimu Wang, Yiwen Jiang, Xiangyu Zhao, Yaling Shen et al.
StratCBT: first large strategy-aligned CBT dataset (9.7K sessions); valuable, but fully LLM-synthetic and GPT-4o-judged, so realism claims need caution.
SAFARI: An Industrial Benchmark for LLM-Assisted Hazard Analysis and Risk Assessment
Chenxi Wu, Zimu Wang, Haiyang Zhang, Wei Wang et al.
First industrial ISO 26262 HARA benchmark (3k real cases, expert-validated LLM judge); LLMs fail ASIL classification (F1 0.261) but stats are thin.
Towards Deployable Underwater Vessel Classification
Abishek Soti, Thura Pyae Sone, Naqib Ibnul, Htoo Htet Aung et al.
Careful, honest benchmark study: recording-level splits cut macro F1 from 0.99 (ShipsEar) to 0.72 (DeepShip); thin statistics and no code release limit impact.
IndicBankBench: Evaluating Safety and Reliability of Language Model Assistants in Indian Retail Banking
Suvradip Paul, Chandra Bhushan, Harsh Sharma, Nitin Kukreja et al.
Shows at-least-once success overstates banking-agent reliability (60–74% vs 43.7–58.2% strict); rigorous staged benchmark, but judge audit is thin.
Edustories: A Collection of Real-world Case Studies from Classroom Practices
Michal \v{S}tef\'anik, Jan Nehyba, Jirina Karasova, Martin Fico et al.
Introduces Edustories, a novel 1,492-case teacher-written dataset; solid resource, but LLM-vs-expert comparison lacks statistical rigor and construct clarity.
Stress-testing Alignment Midtraining
Sid Baines, Jonathan Bostock, Maria Angelica Martinez, Andrew Draganov et al.
Stress-tests show alignment midtraining steers motivation but breaks under 2% conflicting finetuning data; well-scaled yet one seed per cell limits certainty.
Lens: Bringing the Right Semantic Perspective into Focus for Training-Free Multimodal Representation Learning
Xinran Liu, Shouqian Shi, Yixian Chen, Ruizhi Chen et al.
Lens: prompt-only training-free MLLM embeddings, +10.2 P@1 over FreeRet on MMEB; gains ablated and tested, but single benchmark, no CIs, no code.
Predicting Emerging Topics from Outliers: A Prospective Study of Weak Signals in Embedding Space
Evangelia Zve, Gauvain Bourgne, Jean-Gabriel Ganascia
Claims outliers predict emerging topics (F1 0.76–0.97 across protocols); credible, but label circularity and small high-consensus subsets temper it.
The Entropy Triangle Method (ETM): A novel framework for the prevention of cardiac arrhythmia with a review of more than 10,000 patients
Arman daliri
声称通过三角区域过采样器选择在10,646名患者上实现了约85%的窦性/非窦性ECG分类；但选择方法缺乏合理性，且“预防”这一说法言过其实。
To Copy or Not to Copy: Controlling Speculative Decoding via Intrinsic Model Signals
Roy Eisenstadt, Ido Cohen, Edo Cohen-Karlik, Lior Wolf et al.
SwitchSD adds a learned linear copy-intent probe to gate copying vs neural drafting, reporting up to 15% throughput over EAGLE3; gains are consistent but variance is unreported.
When Honesty is Not Enough in AI Debate
Rayne Holland, Liming Zhu, Jason Xue
Formal SIO framework shows debate wins can coexist with hidden information leakage in transcripts; rigorous theory, no LLM experiments, mitigation narrow.
Schema-Anchored Latent Reasoning for Semantic Parsing-Based Knowledge Base Question Answering
Guangze Gao, Zixuan Li, Sikui Zhang, Chunfeng Yuan et al.
Proposes SALR, schema-anchored latent reasoning for KBQA; controlled gains on GrailQA/WebQSP are consistent yet modest, with IID splits declining.
Before the Warning Comes Too Late: Incremental Phone-Scam Detection from Speech
Khang Nhat Hoang Vo, Anh Trac Duc Dinh, Tai Tien Ta, Tho Quan
StreamFraudNet scores scam risk incrementally from raw audio; rigorous and honest, but its global baseline wins on accuracy—incrementality is the only gain.
ASIRF: An Agentic Framework for Context-Dependent Sensitive Information Redaction
Sudha Priyadarshini, Mohamed Chahine Ghanem
Claims KB-at-inference beats trained privacy filter on recall in 85% of 80 combos; results are broad but recall-only, untested for significance, and circular.
Viveka-Insight: a cross-lingual concept graph and citation-grounded retrieval resource over the complete works of Swami Vivekananda in English and Bengali
Tamal Maharaj
Resource paper links non-parallel EN/BN Vivekananda corpora via an LLM concept graph; admirably honest evaluation but core-layer precision is just 0.60.
The Public Discourse Corpus (PDC): A Speaker-Attributed Dataset for Valence and Epistemic Modality with Target Speaker Participation
Bo Chen
Ambitious 186k-sentence speaker-attributed valence+modality corpus with unusually honest validation disclosure; core speaker-attribution accuracy unverified.
Foundations of Stochastic Lexical Calculus: Semantic Descent and Random Dynamics on Probability Simplices
Matthew F Dixon
Formal test for when LLM probabilities can act as a stochastic state; proofs largely standard, gates honest, empirical evidence a small synthetic demo.
EAGER: Enhancing Generative Event Extraction via Reinforcement Learning with Verifiable Rewards
Omar Adjali, Siting Liang, Omair Shahzad Bhatti, Daniel Sonntag
EAGER's decomposed rewards + schema-contrastive RL lift event extraction +5.15 avg F1 over prior RL, but internal table inconsistencies demand scrutiny.
SARFusion: Scene-Aware Routing Fusion for Robust Camera-LiDAR 3D Object Detection
Yuting Zhao, Ziyi Zheng, Shuxiao Li
SARFusion routes queries across camera/LiDAR/fusion branches, reporting gains in fog, rain and snow, but the efficiency cost and corruption protocol are unclear and Figure 6's percentages do not sum to 100.
Fine-Tuning Models for Biomedical Relation Extraction
Claudiu Creanga, Liviu P. Dinu, Daniela Gifu
Claims fine-tuned Gemini 1.0 Pro beats SNPPhenA SOTA (0.89 sentence, 0.80 abstract F1); plausible, but no significance tests or code temper confidence.
Reading Emotions in the Token Space: Discriminative Adaptation of SpeechLLMs for Emotion Recognition
Hasindri Watawana, Sergio Burdisso, Esa\'u Villatoro-Tello, Manjunath K E et al.
Linear head on frozen speechLLM hidden states aids SER on IEMOCAP: consistent small gains and interpretability, but single corpus and no significance tests.
Design of the IBM Granite 5.0 TurboCTC ASR Model
Brian Kingsbury, George Saon, Masayuki Suzuki, Hong-Kwang J. Kuo et al.
Claims a 470M CTC Conformer on the OpenASR Pareto frontier at 2x rival speed; ablations broadly support it, but ~0.1% WER gains lack variance or significance analysis.