Benchmarks are essential for quantitatively tracking progress in AI. As AI agents become increasingly capable, researchers and practitioners have introduced agentic benchmarks to evaluate agents on complex, real-world tasks. These benchmarks typically measure agent capabilities by evaluating task outcomes via specific reward designs. However, we show that many agentic benchmarks have issues in task setup or reward design. For example, SWE-bench-Verified uses insufficient test cases, while $\tau$-bench counts empty responses as successes. Such issues can lead to under- or overestimation of agents’ performance by up to 100% in relative terms. To make agentic evaluation rigorous, we introduce the Agentic Benchmark Checklist (ABC), a set of guidelines that we synthesized from our benchmark-building experience, a survey of best practices, and previously reported issues. When applied to CVE-Bench, a benchmark with a particularly complex evaluation design, ABC reduces performance overestimation by 33%.

Much of the knowledge encoded in transformer language models (LMs) may be expressed in terms of relations: relations between words and their synonyms, entities and their attributes, etc. We show that, for a subset of relations, this computation is well-approximated by a single linear transformation on the subject representation. Linear relation representations may be obtained by constructing a first-order approximation to the LM from a single prompt, and they exist for a variety of factual, commonsense, and linguistic relations. However, we also identify many cases in which LM predictions capture relational knowledge accurately, but this knowledge is not linearly encoded in their representations. Our results thus reveal a simple, interpretable, but heterogeneously deployed knowledge representation strategy in transformer LMs.

This article presents the latest developments to ClaimBuster’s claim-spotting model, which tackles the critical task of identifying check-worthy claims from large streams of information. We introduce the first adversarially regularized, transformer-based claim-spotting model, which achieves state-of-the-art results on several benchmark datasets. In addition to analyzing model performance metrics, we also quantitatively and qualitatively analyze the impact of ClaimBuster’s real-world deployment. Moreover, to help facilitate reproducibility and community engagement, we publicly release our codebase, dataset, data curation platform, API, Google Colab notebooks, and various ClaimBuster-based demo systems, at claimbuster.org .

We analyze the storage and recall of factual associations in autoregressive transformer language models, finding evidence that these associations correspond to localized, directly-editable computations. We first develop a causal intervention for identifying neuron activations that are decisive in a model's factual predictions. This reveals a distinct set of steps in middle-layer feed-forward modules that mediate factual predictions while processing subject tokens. To test our hypothesis that these computations correspond to factual association recall, we modify feed-forward weights to update specific factual associations using Rank-One Model Editing (ROME). We find that ROME is effective on a standard zero-shot relation extraction (zsRE) model-editing task, comparable to existing methods. To perform a more sensitive evaluation, we also evaluate ROME on a new dataset of counterfactual assertions, on which it simultaneously maintains both specificity and generalization, whereas other methods sacrifice one or another. Our results confirm an important role for mid-layer feed-forward modules in storing factual associations and suggest that direct manipulation of computational mechanisms may be a feasible approach for model editing. The code, dataset, visualizations, and an interactive demo notebook are available in the supplemental materials.

Overcoming the visual barrier and developing "see-through vision" has been one of mankind's long-standing dreams. Unlike visible light, Radio Frequency (RF) signals penetrate opaque obstructions and reflect highly off humans. This paper establishes a deep-learning model that can be trained to reconstruct continuous video of a 15-point human skeleton even through visual occlusion. The training process adopts a student/teacher learning procedure inspired by the Feynman learning technique, in which video frames and RF data are first collected simultaneously using a co-located setup containing an optical camera and an RF antenna array transceiver. Next, the video frames are processed with a computer-vision-based gait analysis "teacher" module to generate ground-truth human skeletons for each frame. Then, the same type of skeleton is predicted from corresponding RF data using a "student" deep-learning model consisting of a Residual Convolutional Neural Network (CNN), Region Proposal Network (RPN), and Recurrent Neural Network with Long-Short Term Memory (LSTM) that 1) extracts spatial features from RF images, 2) detects all people present in a scene, and 3) aggregates information over many time-steps, respectively. The model is shown to both accurately and completely predict the pose of humans behind visual obstruction solely using RF signals. Primary academic contributions include the novel many-to-many imaging methodology, unique integration of RPN and LSTM networks, and original training pipeline.

Each year, car accidents on United States roadways claim tens of thousands of lives and injure millions of others, of which almost half involve a combination of two critical pre-crash events: turning and changing lanes. Advanced Driver Assistance Systems (ADAS) that are currently installed in vehicles provide reactive protections that warn drivers of dangers up to 0.5 seconds ahead of collisions. However, rule of thumb suggests 2 seconds for safety in emergency reactions; many lives could be saved even with a slight improvement to the warning time. This paper develops an innovative two-stage neural network model that predicts drivers' actions before fatal collisions can occur. In a novel procedural flow, data is collected from sensors and devices installed inside and outside the vehicle including two cameras, a Global Positioning System (GPS) module, an Onboard Diagnostics-II (OBD-II) interface, and a gyroscope, preprocessed with a Convolutional Neural Network-based (CNN) computer vision model to extract facial movements and rotation, filtered and selected with the Classification and Regression Tree (CART), and modeled with a Recurrent Neural Network w/ Long Short-Term Memory (RNN-LSTM). Results show that the methodology presented in this paper is superior compared to existing ones.