MACEval: A Multi-Agent Continual Evaluation Network for Large Models

arXiv — cs.CV•Thursday, November 13, 2025 at 5:00:00 AM

The introduction of MACEval marks a pivotal advancement in the evaluation of large models, responding to the limitations of existing benchmarks that often suffer from overfitting and data contamination. By utilizing a Multi-Agent Continual Evaluation network, MACEval offers a human-free, scalable solution that can dynamically assess model performance across nine diverse tasks with 23 participating models. This approach not only reduces the data overhead but also integrates seamlessly with existing benchmarks, thereby enhancing the overall credibility and adaptability of model evaluations. The significance of MACEval lies in its potential to keep pace with the rapid advancements in AI, ensuring that evaluations remain relevant and reliable in a field characterized by constant change.

— via World Pulse Now AI Editorial System

Read Original

Was this article worth reading? Share it

Recommended Readings

arXiv — cs.CV2 days ago

Bridging Hidden States in Vision-Language Models

PositiveArtificial Intelligence

Vision-Language Models (VLMs) are emerging models that integrate visual content with natural language. Current methods typically fuse data either early in the encoding process or late through pooled embeddings. This paper introduces a lightweight fusion module utilizing cross-only, bidirectional attention layers to align hidden states from both modalities, enhancing understanding while keeping encoders non-causal. The proposed method aims to improve the performance of VLMs by leveraging the inherent structure of visual and textual data.

Read full article

via arXiv — cs.CV

arXiv — cs.LG2 days ago

Bias-Restrained Prefix Representation Finetuning for Mathematical Reasoning

PositiveArtificial Intelligence

The paper titled 'Bias-Restrained Prefix Representation Finetuning for Mathematical Reasoning' introduces a new method called Bias-REstrained Prefix Representation FineTuning (BREP ReFT). This approach aims to enhance the mathematical reasoning capabilities of models by addressing the limitations of existing Representation finetuning (ReFT) methods, which struggle with mathematical tasks. The study demonstrates that BREP ReFT outperforms both standard ReFT and weight-based Parameter-Efficient finetuning (PEFT) methods through extensive experiments.

Read full article

via arXiv — cs.LG

arXiv — cs.LG2 days ago

Transformers know more than they can tell -- Learning the Collatz sequence

NeutralArtificial Intelligence

The study investigates the ability of transformer models to predict long steps in the Collatz sequence, a complex arithmetic function that maps odd integers to their successors. The accuracy of the models varies significantly depending on the base used for encoding, achieving up to 99.7% accuracy for bases 24 and 32, while dropping to 37% and 25% for bases 11 and 3. Despite these variations, all models exhibit a common learning pattern, accurately predicting inputs with similar residuals modulo 2^p.

Read full article

via arXiv — cs.LG

arXiv — cs.LG2 days ago

Higher-order Neural Additive Models: An Interpretable Machine Learning Model with Feature Interactions

PositiveArtificial Intelligence

Higher-order Neural Additive Models (HONAMs) have been introduced as an advancement over Neural Additive Models (NAMs), which are known for their predictive performance and interpretability. HONAMs address the limitation of NAMs by effectively capturing feature interactions of arbitrary orders, enhancing predictive accuracy while maintaining interpretability, crucial for high-stakes applications. The source code for HONAM is publicly available on GitHub.

Read full article

via arXiv — cs.LG