Singapore

Reports of “emergent” capabilities in transformer-based LLMs (abrupt, non-linear improvements in task performance) remain controversial due to post-hoc measurement and ambiguous definitions. We investigate whether such transitions can be predicted pre-hoc from internal representations. For each training checkpoint, we train sparse autoencoders (SAEs) on model activations, construct a co-activation graph over SAE features, and track graph statistics (e.g. density, clustering, etc.). We analyze a 2-layer grokking-style transformer with aggregate graph metrics over eight SAE initializations, and test lead-lag relationships between changes in graph metrics and subsequent changes in accuracy under a formalized emergence criterion. Across both settings, we find no statistically significant evidence that global co-activation topology forecasts emergent jumps in performance. If pre-hoc indicators exist, they may lie outside the global graph measures analyzed here (e.g., in task-specific circuits or localized subgraphs).

AAAI 2026

Predicting Emergent Capabilities Using Sparse Features

transformer training dynamics

grokking

representational topology

co-activation graphs

emergent capabilities

sparse autoencoders

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

ACL 2025

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

workshop paper

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-26 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to AAAI 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://aaai.org/conference/aaai/aaai-26/registration/) first.

**Online Registration Form**: https://aaai.getregistered.net/conference-2026 

Registration Required

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

We present a differentiable framework that leverages the Discrete Empirical Interpolation Method (DEIM) for interpretable deep learning and dynamical system analysis. Although DEIM efficiently approximates nonlinear terms in projection-based reduced-order models (POD-ROM), its fixed interpolation points limit the adaptability to complex and time-varying dynamics. To address this limitation, we first develop a differentiable adaptive DEIM formulation for the one-dimensional viscous Burgers’ equation, which allows neural networks to dynamically select interpolation points in a computationally efficient and physically consistent manner. We then apply DEIM as an interpretable analysis tool for examining the learned dynamics of a pre-trained Neural Ordinary Differential Equation (NODE) on a two-dimensional vortex-merging problem. The DEIM trajectories reveal physically meaningful features in the NODE’s learned dynamics and expose its limitations when extrapolating to unseen flow configurations. These findings demonstrate that DEIM can serve not only as a model reduction tool but also as a diagnostic framework for understanding and improving the generalization behavior of neural differential equation models.

Towards Interpretable Deep Learning and Analysis of Dynamical Systems via the Discrete Empirical Interpolation Method

Classifying non-Hermitian energy spectra under open boundary conditions is an open challenge in physics. This classification is a critical prerequisite for the rational inverse design of systems exhibiting desired dynamics and topological responses. While graph topology has emerged as a promising approach for characterizing these spectra, systematic methods for distilling non-Hermitian spectra into their corresponding graph representations have been lacking. Moreover, the resulting graphs often exhibit complexities that defy manual classification, necessitating machine learning approaches. In this work, we introduce a two-step framework for classifying non-Hermitian spectra based on their graph topologies. The first step employs $\texttt{Poly2Graph}$, an automated, high-performance pipeline that distills non-Hermitian spectra into $\textit{spectral graphs}$ suitable for graph neural networks (GNNs). The second step involves generating a large dataset of these spectral graphs and training a GNN for classification. We propose $\texttt{GnLTransformer}$, a novel architecture featuring dual channels that leverage line graphs to explicitly capture higher-order topological features. $\texttt{GnLTransformer}$ achieves over 99% classification accuracy on our dataset, outperforming standard baselines by 32%. Notably, beyond conventional GNNs, $\texttt{GnLTransformer}$ offers inherent explainability regarding higher-order topology. As a further contribution, we release a new $\textit{multi-graph}$ dataset comprising over 117K spectral graphs.

Classifying the Graph Topology of Non-Hermitian Energy Spectra with Graph Transformer

Machine learning force fields (MLFFs) now rival quantum methods in accuracy, yet their internal logic remains opaque—a critical barrier to both trust and scientific discovery. We introduce **DUAL-X**, a *Dual-Level Explainability Framework* that bridges model reasoning with human understanding. **DUAL-X** unites two complementary perspectives: a **model-centric** level identifying *where* in the atomic structure the model focuses its attention, and a **human-centric** level revealing *what* physically meaningful interactions it prioritizes. Implemented with Grad-CAM for spatial attribution and SHAP-on-SOAP for physical interpretation, **DUAL-X** provides a general, human-in-the-loop paradigm for interpretable scientific AI. Applied to dopant migration—a challenging task for MLFFs—**DUAL-X** reveals that training strategy governs the emergence of chemical intuition. Multi-temperature fine-tuned models exhibit over $10^2\times$ stronger selectivity for Cr–Cr $f$-type ($l=3$) angular correlations than scratch-trained ones, emphasizing complex 3D coordination motifs essential for accuracy. Models with sharper Grad-CAM focus also display coherent SHAP importance for dopant clustering, revealing consistent internal reasoning across scales. By aligning model logic with human physical knowledge, **DUAL-X** transforms opaque predictors into interpretable scientific partners—advancing trustworthy, explainable, and insight-driven AI for materials discovery.

What is Your Force Field Really Learning? Gaining Scientific Intuition with A Dual-Level Explainability Framework

In data-scarce settings, learning accurate and interpretable models for high-stakes medical tabular classification remains a fundamental challenge, as healthcare decisions must be transparent and trustworthy. We propose a novel pipeline for few-shot tabular classification that distills the predictive power of large foundation models into globally interpretable student models. Given a tabular dataset, we generate synthetic data using CTGAN to approximate the underlying distribution. We then train high-capacity teacher models (TabPFN, TabM) on a small number of labeled examples and use them to pseudolabel the synthetic data. Finally, we train student explainer models (i.e., XGBoost, decision trees, Generalized Linear Rule Model (GLRM), and TTnet) on this pseudolabeled synthetic dataset. These student models are globally and exactly interpretable, yielding logical decision rules (e.g., disjunctive normal form) that fully reproduce their predictions. Evaluated across 8 clinical tabular tasks, our distilled models generally outperform baselines trained directly on the few-shot data, with improved ROC AUC scores across few-shot settings. This work demonstrates that foundation models can be effectively leveraged as teachers to produce small, transparent, and high-performing classifiers. Our approach advances the goal of reliable and interpretable machine learning in real-world settings where labeled data is limited.

Leveraging Foundation Models in Healthcare: A Distillation Approach to Interpretable Clinical Prediction

Adjoint-based inverse design yields compact, high-performance nanophotonic devices, but the mapping from pixel-level layouts to optical figures of merit remains hard to interpret. We present a simple pipeline that (i) generates a large set of wavelength demultiplexers (WDMs) with SPINS-B, (ii) records each final 2D layout and its spectral metrics (e.g., transmitted power at 1310 nm and 1550 nm), and (iii) trains a lightweight convolutional surrogate to predict these metrics from layouts, enabling (iv) gradient-based attribution via Integrated Gradients (IG) to highlight specific regions most responsible for performance. On a corpus of sampled WDMs, IG saliency consistently localizes to physically meaningful features (e.g., tapers and splitter hubs), offering design intuition that complements adjoint optimization. Our contribution is an end-to-end, data-driven workflow—SPINS-B dataset, CNN surrogate, and IG analysis—that turns inverse-designed layouts into interpretable attributions without modifying the physics solver or objective, and that can be reused for other photonic components.

Towards Explainable Inverse Design for Photonics via Integrated Gradients

Recent research has extensively studied how large language models manipulate integers in specific arithmetic tasks, and on a more fundamental level, how they represent numeric values. These previous works have found that language model embeddings can be used to reconstruct the original values, however, they do not evaluate whether language models actually model continuous values as continuous. Using expected properties of the embedding space, including linear reconstruction and principal component analysis, we show that language models not only represent numeric spaces as non-continuous but also introduce significant noise. Using models from three major providers (OpenAI, Google Gemini and Voyage AI), we show that while reconstruction is possible with high fidelity ($R^2 \geq 0.95$), principal components only explain a minor share of variation within the embedding space. This indicates that many components within the embedding space are orthogonal to the simple numeric input space. Further, both linear reconstruction and explained variance suffer with increasing decimal precision, despite the ordinal nature of the input space being fundamentally unchanged. The findings of this work therefore have implications for the many areas where embedding models are used, in-particular where high numerical precision, large magnitudes or mixed-sign values are common.

Language Models Do Not Embed Numbers Continuously

Cell phenotype transition refers to the changes in the morphology, function, and surface markers of cells that occur under specific environmental conditions or physiological states, based on their genomic information and external signals. This process plays an important role in development, tissue repair, and responses to external stimuli such as infection or inflammation. Traditional bioinformatics methods for addressing cell type transition often rely on hypothesis-driven models, which may not fully capture the complexity and heterogeneity of the transition processes. In this paper, we introduce DualCPT, a cell phenotype transition and differentiation model based on Markov processes. Specifically, the model consists of a classification branch and a transition branch. The transition branch identifies regulatory genes involved in cell phenotype transition and differentiation. In the classification branch, we evaluate the model’s overall performance on general cell type classification tasks using a comprehensive multi-metric evaluation framework; in the transition branch, we implement a token pruning-based approach for critical locus discovery and enhance information interaction between full-sequence contexts and prioritized regulatory sites via an improved multi-head attention mechanism. Cell phenotype transition tasks are further assessed by uncertainty quantification and confidence calibration. In particular, in gene knockout experiments, we found that knocking out important genes alters the probability of cell phenotype transition and differentiation, and knocking out a certain number of essential genes can terminate specific transition processes. Data, code, and checkpoints are publicly available at https://github.com/Ssupercoder/DualCPT.

DualCPT: Dual-branch Modeling for Cellular Phenotype Transition

In recent years, the challenge of handling out-of-distribution (OOD) data has become central to ensuring the reliability of artificial intelligence deployed in open-world environments. This half-day tutorial at AAAI’26 is designed for students, researchers, and practitioners seeking to advance the machine learning models by providing an integrated overview of OOD detection and OOD generalization. The tutorial will introduce foundational concepts and theoretical principles, survey advanced methodologies across post-hoc scoring, outlier exposure, representation learning, as well as causality-inspired approaches, and highlight challenges together with emerging solutions. Real-world case studies across different domains will demonstrate the critical importance of addressing OOD risks in practice, while discussions will emphasize the distinctions and synergies between detection and generalization as well as open research directions. To participate effectively, attendees are expected to have basic familiarity with linear algebra, probability, and fundamental machine learning concepts, though the content is designed to remain accessible and emphasize intuitive understanding over technical formalism.

Handling Out-of-Distribution Data in the Open World: Principles and Practice for Reliable AI

Although many LLM-based agents have recently been developed, progress on agents that are capable of solving complex, real-world problems is severely limited. One such important and challenging area is that of addressing IT management tasks, including solving IT incidents, which often require extensive human expertise and effort. To enable the development of agents for these tasks, in this lab, we introduce IT-Bench, an open benchmark for IT automation that simulates realistic environments where agents interact with IT systems and multi-modal operational data, including logs, metrics, alerts, and traces. ITBench provides a two- tiered benchmark: a static dataset (ITBench_static) that closely mirrors the live, gym-like environment (ITBench_live). Together, they provide a testbed for benchmarking agentic systems across a host of critical challenges, such as planning and reasoning over massive and heterogeneous IT data, safety, and stochasticity in live IT systems. Through demonstrations, participants will learn all of these challenges. ITBench_static is more beginner-friendly, enabling rapid benchmarking of agents against a smaller set of unique IT-domain challenges. ITBench_live enables more advanced researchers and practitioners familiar with IT systems to develop and test their agents against the full suite of challenging IT problems. We will guide attendees to use both ITBench_static and ITBench_live to develop baseline multi-agent systems and benchmark their performance. No prior IT or site reliability engineering (SRE) experience is required.

Downloads

Next from AAAI 2026

Towards Interpretable Deep Learning and Analysis of Dynamical Systems via the Discrete Empirical Interpolation Method

Similar lecture

Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAAI 2026

Towards Interpretable Deep Learning and Analysis of Dynamical Systems via the Discrete Empirical Interpolation Method

Similar lecture

Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads