Adam Byerly

Kevin Xu, Yeganeh Kordi, Tanay Nayak et al. · uw

Can advanced multi-modal models effectively tackle complex web-based tasks? Such tasks are often found on crowdsourcing platforms, where crowdworkers engage in challenging micro-tasks within web-based environments. Building on this idea, we present TurkingBench, a benchmark consisting of tasks presented as web pages with textual instructions and multi-modal contexts. Unlike previous approaches that rely on artificially synthesized web pages, our benchmark uses natural HTML pages originally designed for crowdsourcing workers to perform various annotation tasks. Each task's HTML instructions are instantiated with different values derived from crowdsourcing tasks, creating diverse instances. This benchmark includes 32.2K instances spread across 158 tasks. To support the evaluation of TurkingBench, we have developed a framework that links chatbot responses to actions on web pages (e.g., modifying a text box, selecting a radio button). We assess the performance of cutting-edge private and open-source models, including language-only and vision-language models (such as GPT4 and InternVL), on this benchmark. Our results show that while these models outperform random chance, there is still significant room for improvement. We hope that this benchmark will drive progress in the evaluation and development of web-based agents.

Adam Byerly, Daniel Khashabi

Large language models exhibit a strong position bias in multi-document contexts, systematically prioritizing information based on location rather than relevance. While existing approaches treat this bias as noise to be mitigated, we introduce Gold Panning Bandits, a framework that leverages position bias as a diagnostic signal: by reordering documents and observing shifts in the model's responses, we can efficiently identify the most relevant content. We frame the problem of choosing reorderings as a bipartite matching problem. While an optimal assignment can be computed at each iteration with the Hungarian algorithm in $O(N^3)$ time, we propose a greedy $O(N \log N)$ strategy that achieves comparable performance by prioritizing the placement of the most uncertain documents in the most informative positions. Our approach identifies relevant documents using up to 65\% fewer language model queries than random permutation baselines on knowledge-intensive NLP tasks, substantially reducing computational cost without model retraining. This work demonstrates that inherent LLM biases can be transformed from liabilities into assets for efficient, inference-time optimization.

Adam Byerly, Daniel Khashabi

Self-consistency (SC) improves the performance of large language models (LLMs) across various tasks and domains that involve short content. However, does this support its effectiveness for long-context problems? We challenge the assumption that SC's benefits generalize to long-context settings, where LLMs often struggle with position bias, the systematic over-reliance on specific context regions-which hinders their ability to utilize information effectively from all parts of their context. Through comprehensive experimentation with varying state-of-the-art models, tasks, and SC formulations, we find that SC not only fails to improve but actively degrades performance on long-context tasks. This degradation is driven by persistent position bias, which worsens with longer context lengths and smaller model sizes but remains invariant to prompt format or task type. Unlike short-context tasks, where SC diversifies reasoning paths, long-context SC amplifies positional errors. These comprehensive results provide valuable insight into the limitations of current LLMs in long-context understanding and highlight the need for more sophisticated approaches.

Taiming Lu, Muhan Gao, Kuai Yu et al.

Large Language Models (LLMs) exhibit positional bias, struggling to utilize information from the middle or end of long contexts. Our study explores LLMs' long-context reasoning by probing their hidden representations. We find that while LLMs encode the position of target information, they often fail to leverage this in generating accurate responses. This reveals a disconnect between information retrieval and utilization, a "know but don't tell" phenomenon. We further analyze the relationship between extraction time and final accuracy, offering insights into the underlying mechanics of transformer models.

Adam Byerly, Tatiana Kalganova

We provide a definition for class density that can be used to measure the aggregate similarity of the samples within each of the classes in a high-dimensional, unstructured dataset. We then put forth several candidate methods for calculating class density and analyze the correlation between the values each method produces with the corresponding individual class test accuracies achieved on a trained model. Additionally, we propose a definition for dataset quality for high-dimensional, unstructured data and show that those datasets that met a certain quality threshold (experimentally demonstrated to be > 10 for the datasets studied) were candidates for eliding redundant data based on the individual class densities.

Adam Byerly, Tatiana Kalganova

We show that, for each of five datasets of increasing complexity, certain training samples are more informative of class membership than others. These samples can be identified a priori to training by analyzing their position in reduced dimensional space relative to the classes' centroids. Specifically, we demonstrate that samples nearer the classes' centroids are less informative than those that are furthest from it. For all five datasets, we show that there is no statistically significant difference between training on the entire training set and when excluding up to 2% of the data nearest to each class's centroid.

Adam Byerly, Tatiana Kalganova, Anthony J. Grichnik

We present a dataset consisting of high-resolution images of 13 micro-PCBs captured in various rotations and perspectives relative to the camera, with each sample labeled for PCB type, rotation category, and perspective categories. We then present the design and results of experimentation on combinations of rotations and perspectives used during training and the resulting impact on test accuracy. We then show when and how well data augmentation techniques are capable of simulating rotations vs. perspectives not present in the training data. We perform all experiments using CNNs with and without homogeneous vector capsules (HVCs) and investigate and show the capsules' ability to better encode the equivariance of the sub-components of the micro-PCBs. The results of our experiments lead us to conclude that training a neural network equipped with HVCs, capable of modeling equivariance among sub-components, coupled with training on a diversity of perspectives, achieves the greatest classification accuracy on micro-PCB data.

Adam Byerly, Tatiana Kalganova, Ian Dear

Most capsule network designs rely on traditional matrix multiplication between capsule layers and computationally expensive routing mechanisms to deal with the capsule dimensional entanglement that the matrix multiplication introduces. By using Homogeneous Vector Capsules (HVCs), which use element-wise multiplication rather than matrix multiplication, the dimensions of the capsules remain unentangled. In this work, we study HVCs as applied to the highly structured MNIST dataset in order to produce a direct comparison to the capsule research direction of Geoffrey Hinton, et al. In our study, we show that a simple convolutional neural network using HVCs performs as well as the prior best performing capsule network on MNIST using 5.5x fewer parameters, 4x fewer training epochs, no reconstruction sub-network, and requiring no routing mechanism. The addition of multiple classification branches to the network establishes a new state of the art for the MNIST dataset with an accuracy of 99.87% for an ensemble of these models, as well as establishing a new state of the art for a single model (99.83% accurate).

Adam Byerly, Tatiana Kalganova

Capsules are the name given by Geoffrey Hinton to vector-valued neurons. Neural networks traditionally produce a scalar value for an activated neuron. Capsules, on the other hand, produce a vector of values, which Hinton argues correspond to a single, composite feature wherein the values of the components of the vectors indicate properties of the feature such as transformation or contrast. We present a new way of parameterizing and training capsules that we refer to as homogeneous vector capsules (HVCs). We demonstrate, experimentally, that altering a convolutional neural network (CNN) to use HVCs can achieve superior classification accuracy without increasing the number of parameters or operations in its architecture as compared to a CNN using a single final fully connected layer. Additionally, the introduction of HVCs enables the use of adaptive gradient descent, reducing the dependence a model's achievable accuracy has on the finely tuned hyperparameters of a non-adaptive optimizer. We demonstrate our method and results using two neural network architectures. First, a very simple monolithic CNN that when using HVCs achieved a 63% improvement in top-1 classification accuracy and a 35% improvement in top-5 classification accuracy over the baseline architecture. Second, with the CNN architecture referred to as Inception v3 that achieved similar accuracies both with and without HVCs. Additionally, the simple monolithic CNN when using HVCs showed no overfitting after more than 300 epochs whereas the baseline showed overfitting after 30 epochs. We use the ImageNet ILSVRC 2012 classification challenge dataset with both networks.