arXiv subjects

Ben Slater

Publications and source records attributed to Ben Slater.

At least 19 recordsLinked to original sources

arXiv2026

Elucidating Guest-Host Mechanisms in ZIF-L for Tuneable Highly Luminescent 2D Materials

3D metal-organic frameworks (MOFs) are well known effective hosts for luminescent guests that improve tuneability, photostability and material fabricability. 2D guest@MOF systems, while less explored, offer competitive advantages over 3D guest@MOF systems due to their optical transparency, inter-layer spacing, and vertical thinness. This work examines luminescent organic dyes@ZIF-L to establish the underlying mechanisms of guest incorporation in ZIF-L and demonstrate the advantages of the 2D ZIF-L architecture for the material's functional luminescence. Analysing a case study system, fluorescein@ZIF-L (F@ZIF-L), using nanoscale FTIR, diffraction, and topology mapping, confirmed that fluorescein (F) resided in the ZIF-L framework cavities. Supported by surface energy simulations, the extent of guest incorporation was found to be indicated by a morphological continuum, from the characteristic leaf-shaped ZIF-L to rectangular F@ZIF-L. By modifying synthesis temperature and solvent ratios, the luminescent properties of F@ZIF-L could be rationally tuned in terms of guest loading (% mol) and arrangement (i.e. guest monomer to aggregate ratio). When optimised, F@ZIF-L exhibited tuneable emission chromaticity, 99.7% photoluminescent quantum yield, minimal guest leaching in solution over 12 months, and high photostability. Perylene@ZIF-L exhibited unique white light emitting properties, with CIE coordinates (0.33, 0.34) arising from a combination of yellow alpha-phase excimer and blue monomeric perylene emission. Finally, oriented luminescent thin films of guest@ZIF-L materials were grown on malleable Zn foils, demonstrating an in situ fabrication technique. Together, the work highlights the potential for luminescent dye@ZIF-L systems in developing tuneable and resilient luminescent components of next-generation optoelectronics, sensors, and lighting systems.

cond-mat.mtrl-sci↗

arXiv2026

Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action

Theory of Mind (ToM) benchmarks for Large Language Models (LLMs) typically rely on passive question-answering formats, but the deployment of LLMs in increasingly agentic and autonomous forms demands new evaluations. In this paper we evaluate an agent's ability to induce specific belief states in other agents by taking actions rather than using conversational persuasion, a capability we call Non-Conversational Planning ToM (NCP-ToM). NCP-ToM is likely to be essential for many agent use-cases, including within user-assistant interactions and pedagogical contexts, but may also present manipulation or misinformation risks. Using a novel framework, NCP-ExploreToM, we subvert the conventional task structure by providing models with a set of belief state goals and requiring them to move objects or direct characters into rooms to achieve their goals. We evaluated six frontier models, including GPT-5, Gemini 2.5 Pro and the Claude 4 series, and a cohort of human participants, across 600 task instances. GPT-5 was successful on approximately 80% of tasks in the agentic setting, and was the only model to outperform human participants on our task, but was still less robust than humans across contexts. We additionally found that all models, like humans, performed better on tasks inducing true belief states than false belief states, which is a positive signal for alignment efforts. These findings highlight emerging social-reasoning capabilities in LLMs for non-conversational task completion and underscore the necessity of agentic evaluations for understanding the safety and alignment of autonomous social agents.

cs.CL↗

arXiv2026

Capabilities Ain't All You Need: Measuring Propensities in AI

AI evaluation has primarily focused on measuring capabilities, with formal approaches inspired from Item Response Theory (IRT) being increasingly applied. Yet propensities - the tendencies of models to exhibit particular behaviours - play a central role in determining both performance and safety outcomes. However, traditional IRT describes a model's success on a task as a monotonic function of model capabilities and task demands, an approach unsuited to propensities, where both excess and deficiency can be problematic. Here, we introduce the first formal framework for measuring AI propensities by using a bilogistic formulation for model success, which attributes high success probability when the model's propensity is within an "ideal band". Further, we estimate the limits of the ideal band using LLMs equipped with newly developed task-agnostic rubrics. Applying our framework to six families of LLM models whose propensities are incited in either direction, we find that we can measure how much the propensity is shifted and what effect this has on the tasks. Critically, propensities estimated using one benchmark successfully predict behaviour on held-out tasks. Moreover, we obtain stronger predictive power when combining propensities and capabilities than either separately. More broadly, our framework showcases how rigorous propensity measurements can be conducted and how it yields gains over solely using capability evaluations to predict AI behaviour.

cs.LG↗

arXiv2025

I Spy With My Model's Eye: Visual Search as a Behavioural Test for MLLMs

Multimodal large language models (MLLMs) achieve strong performance on vision-language tasks, yet their visual processing is opaque. Most black-box evaluations measure task accuracy, but reveal little about underlying mechanisms. Drawing on cognitive psychology, we adapt classic visual search paradigms -- originally developed to study human perception -- to test whether MLLMs exhibit the ``pop-out'' effect, where salient visual features are detected independently of distractor set size. Using controlled experiments targeting colour, size and lighting features, we find that advanced MLLMs exhibit human-like pop-out effects in colour or size-based disjunctive (single feature) search, as well as capacity limits for conjunctive (multiple feature) search. We also find evidence to suggest that MLLMs, like humans, incorporate natural scene priors such as lighting direction into object representations. We reinforce our findings using targeted fine-tuning and mechanistic interpretability analyses. Our work shows how visual search can serve as a cognitively grounded diagnostic tool for evaluating perceptual capabilities in MLLMs.

cs.CV↗

arXiv2025

PredictaBoard: Benchmarking LLM Score Predictability

Despite possessing impressive skills, Large Language Models (LLMs) often fail unpredictably, demonstrating inconsistent success in even basic common sense reasoning tasks. This unpredictability poses a significant challenge to ensuring their safe deployment, as identifying and operating within a reliable "safe zone" is essential for mitigating risks. To address this, we present PredictaBoard, a novel collaborative benchmarking framework designed to evaluate the ability of score predictors (referred to as assessors) to anticipate LLM errors on specific task instances (i.e., prompts) from existing datasets. PredictaBoard evaluates pairs of LLMs and assessors by considering the rejection rate at different tolerance errors. As such, PredictaBoard stimulates research into developing better assessors and making LLMs more predictable, not only with a higher average performance. We conduct illustrative experiments using baseline assessors and state-of-the-art LLMs. PredictaBoard highlights the critical need to evaluate predictability alongside performance, paving the way for safer AI systems where errors are not only minimised but also anticipated and effectively mitigated. Code for our benchmark can be found at https://github.com/Kinds-of-Intelligence-CFI/PredictaBoard

cs.CL↗

arXiv2024

A little less conversation, a little more action, please: Investigating the physical common-sense of LLMs in a 3D embodied environment

As general-purpose tools, Large Language Models (LLMs) must often reason about everyday physical environments. In a question-and-answer capacity, understanding the interactions of physical objects may be necessary to give appropriate responses. Moreover, LLMs are increasingly used as reasoning engines in agentic systems, designing and controlling their action sequences. The vast majority of research has tackled this issue using static benchmarks, comprised of text or image-based questions about the physical world. However, these benchmarks do not capture the complexity and nuance of real-life physical processes. Here we advocate for a second, relatively unexplored, approach: 'embodying' the LLMs by granting them control of an agent within a 3D environment. We present the first embodied and cognitively meaningful evaluation of physical common-sense reasoning in LLMs. Our framework allows direct comparison of LLMs with other embodied agents, such as those based on Deep Reinforcement Learning, and human and non-human animals. We employ the Animal-AI (AAI) environment, a simulated 3D virtual laboratory, to study physical common-sense reasoning in LLMs. For this, we use the AAI Testbed, a suite of experiments that replicate laboratory studies with non-human animals, to study physical reasoning capabilities including distance estimation, tracking out-of-sight objects, and tool use. We demonstrate that state-of-the-art multi-modal models with no finetuning can complete this style of task, allowing meaningful comparison to the entrants of the 2019 Animal-AI Olympics competition and to human children. Our results show that LLMs are currently outperformed by human children on these tasks. We argue that this approach allows the study of physical reasoning using ecologically valid experiments drawn directly from cognitive science, improving the predictability and reliability of LLMs.

cs.AI↗

arXiv2023

Enhanced elastic stability of a topologically disordered crystalline metal--organic framework

By virtue of their open network structures and low densities, metal--organic frameworks (MOFs) are soft materials that exhibit elastic instabilities at low applied stresses. The conventional strategy for improving elastic stability is to increase the connectivity of the underlying MOF network, which necessarily increases material density and reduces porosity. Here we demonstrate an alternative paradigm, whereby elastic stability is enhanced in a MOF with an aperiodic network topology. We use a combination of variable-pressure single-crystal X-ray diffraction measurements and coarse-grained lattice-dynamical calculations to interrogate the high-pressure behaviour of the topologically aperiodic system TRUMOF-1, which we compare against that of its ordered congener MOF-5. We show that the topology of the former quenches the elastic instability responsible for pressure-induced framework collapse in the latter, much as irregularity in the shapes and sizes of stones acts to prevent cooperative mechanical failure in drystone walls. Our results establish aperiodicity as a counterintuitive design motif in engineering the mechanical properties of framework structures, relevant to MOFs and larger-scale architectures alike.

cond-mat.mtrl-sci↗

arXiv2023

The Animal-AI Environment: A Virtual Laboratory For Comparative Cognition and Artificial Intelligence Research

The Animal-AI Environment is a unique game-based research platform designed to facilitate collaboration between the artificial intelligence and comparative cognition research communities. In this paper, we present the latest version of the Animal-AI Environment, outlining several major features that make the game more engaging for humans and more complex for AI systems. These features include interactive buttons, reward dispensers, and player notifications, as well as an overhaul of the environment's graphics and processing for significant improvements in agent training time and quality of the human player experience. We provide detailed guidance on how to build computational and behavioural experiments with the Animal-AI Environment. We present results from a series of agents, including the state-of-the-art deep reinforcement learning agent Dreamer-v3, on newly designed tests and the Animal-AI Testbed of 900 tasks inspired by research in the field of comparative cognition. The Animal-AI Environment offers a new approach for modelling cognition in humans and non-human animals, and for building biologically inspired artificial intelligence.

cs.AI↗

arXiv2023

Structure and dynamics of the negative thermal expansion material Cd(CN)$_2$ under pressure

We use a combination of variable-temperature / variable-pressure neutron powder diffraction, variable-pressure inelastic neutron scattering, and quantum chemical calculations to interrogate the behaviour of the negative thermal expansion (NTE) material $^{114}$Cd(CN)$_2$ under hydrostatic pressure. We determine the equation of state of the ambient-pressure phase, and discover the so-called `warm hardening' effect whereby the material becomes elastically stiffer as it is heated. We also identify a number of high-pressure phases, and map out the phase behaviour of Cd(CN)$_2$ over the range $0\leq p\leq0.5$\,GPa, $100\leq T\leq300$\,K. As expected for an NTE material, the low-energy phonon frequencies are found to soften under pressure, and we determine an effective Gr{\"u}neisen parameter for these modes. Finally, we show that the elastic behaviour of Cd(CN)$_2$ is sensitive to the local Cd coordination environment, which suggests an interplay between short- (phononic) and long-timescale (cyanide flips) fluctuations in Cd(CN)$_2$.

cond-mat.mtrl-sci↗

arXiv2022

Truchet-tile structure of a topologically aperiodic metal-organic framework

Periodic tilings can store information if individual tiles are decorated to lower their symmetry. Truchet tilings - the broad family of space-filling arrangements of such tiles - offer an efficient mechanism of visual data storage related to that used in barcodes and QR codes. Here, we show that the crystalline metal-organic framework [OZn$_4$][1,3-benzenedicarboxylate]$_3$ (TRUMOF-1) is an atomic-scale realisation of a complex three-dimensional Truchet tiling. Its crystal structure consists of a periodically-arranged assembly of identical zinc-containing clusters connected uniformly in a well-defined but disordered fashion to give a topologically aperiodic microporous network. We suggest that this unusual structure emerges as a consequence of geometric frustration in the chemical building units from which it is assembled.

cond-mat.mtrl-sci↗

arXiv2019

Room temperature spin-ice physics in cadmium cyanide

Spin-ices are frustrated magnets that support a particularly rich variety of emergent physics. Typically, it is the interplay of magnetic dipole interactions, spin anisotropy, and geometric frustration on the pyrochlore lattice that drives spin-ice formation. The relevant physics occurs at temperatures commensurate with the magnetic interaction strength, which for most systems is 1--5\,K. This low energy scale poses severe challenges for experimental studies of spin-ices and the practical exploitation of their unusual properties. Here, we show that non-magnetic cadmium cyanide (Cd(CN)$_2$) exhibits analogous behaviour to magnetic spin-ices, but does so on a temperature scale that is nearly two orders of magnitude greater. The electric dipole moments of cyanide ions in Cd(CN)$_2$ assume the role of magnetic pseudospins, with the difference in energy scale reflecting the increased strength of electric \emph{vs} magnetic dipolar interactions. As a result, spin-ice physics influences the structural behaviour of Cd(CN)$_2$ even at room temperature.

cond-mat.mtrl-sci↗

arXiv2017

Is High-density Amorphous Ice Simply a 'Derailed' State along the Ice I to Ice IV Pathway?

The structural nature of high-density amorphous ice (HDA), which forms through low-temperature pressure-induced amorphization of the 'ordinary' ice I, is heavily debated. Clarifying this question is not only important for understanding the complex condensed states of H$_2$O but also in the wider context of pressure-induced amorphization processes, which are encountered across the entire materials spectrum. We first show that ammonium fluoride (NH$_4$F), which has a similar hydrogen-bonded network to ice I, also undergoes a pressure collapse upon compression at 77 K. However, the product material is not amorphous but NH$_4$F II, a high-pressure phase isostructural with ice IV. This collapse can be rationalized in terms of a highly effective mechanism. In the case of ice I, the orientational disorder of the water molecules leads to a deviation from this mechanism and we therefore classify HDA as a 'derailed' state along the ice I to ice IV pathway.

cond-mat.mtrl-sci↗

arXiv2016

Violations of L\"owensteins rule in zeolites

Zeolites, microporous aluminosilicates, are amongst the most widely used catalysts in the petrochemical industry. Zeolite catalytic functionality is coupled to the distribution of tetrahedral alumina (AlO4-) and associated counter-cations throughout the aluminosilicate framework, yet little is definitively known about the factors that govern framework aluminium arrangement. It is generally accepted that all zeolites obey Lowensteins rule of aluminium avoidance and that Al-O-Al linkages are forbidden. Here, we describe the unprecedented screening of aluminium distribution in catalytically active zeolite SSZ-13 in both its protonated and sodium containing forms, H-SSZ-13 and Na-SSZ-13, using density functional theory. We predict violations of Lowensteins rule in high and low silica H-SSZ-13 and other protonated frameworks considered in this investigation, H-LTA, H-RHO and H-ABW. The synthetic realisation of these zeolites could spur the development of new catalytic routes and materials, and the optimisation of existing zeolite catalysts.

cond-mat.mtrl-sci↗

arXiv2016

Detailed crystallographic analysis of the ice VI to ice XV hydrogen ordering phase transition

The D2O ice VI to ice XV hydrogen ordering phase transition at ambient pressure is investigated in detail with neutron diffraction. The lattice constants are found to be sensitive indicators for hydrogen ordering. The a and b lattice constants contract whereas a pronounced expansion in c is found upon hydrogen ordering. Overall, the hydrogen ordering transition goes along with a small increase in volume which explains why the phase transition is more difficult to observe upon cooling under pressure. Slow-cooling ice VI at 1.4 GPa gives essentially fully hydrogen-disordered ice VI. Consistent with earlier studies, the ice XV obtained after slow-cooling at ambient pressure is best described with P-1 space group symmetry. Using a new computational approach, we achieve the atomistic reconstruction of a supercell structure that is consistent with the average partially ordered structure derived from Rietveld refinements. This shows that C-type networks are most prevalent in ice XV but other structural motifs outside of the classifications of the fully hydrogen-ordered networks are identified as well. The recently proposed Pmmn structural model for ice XV is found to be incompatible with our diffraction data and we argue that only structural models that are capable of describing full hydrogen order should be used.

cond-mat.mtrl-sci↗

arXiv2015

Dynamic Acidity in Defective UiO-66

The metal organic framework (MOF) material UiO-66 has emerged as one of the most promising MOF materials due to its thermal and chemical stability and its potential for catalytic applications. Typically, as-synthesised UiO-66 has a relatively high concentration of missing linker defects. The presence of these defects has been correlated with catalytic activity but characterisation of defect structure has proved elusive. We refine a recent experimental determination of defect structure using static and dynamic first principles approaches, which reveals a dynamic and labile acid centre that could be tailored for functional applications in catalysis.

cond-mat.mtrl-sci↗

arXiv2015

Molecular simulations of heterogeneous ice nucleation. I. Controlling ice nucleation through surface hydrophilicity

Ice formation is one of the most common and important processes on earth and almost always occurs at the surface of a material. A basic understanding of how the physicochemical properties of a material's surface affect its ability to form ice has remained elusive. Here, we use molecular dynamics simulations to directly probe heterogeneous ice nucleation at a hexagonal surface of a nanoparticle of varying hydrophilicity. Surprisingly, we find that structurally identical surfaces can both inhibit and promote ice formation and analogous to a chemical catalyst, it is found that an optimal interaction between the surface and the water exists for promoting ice nucleation. We use our microscopic understanding of the mechanism to design a modified surface in silico with enhanced ice nucleating ability.

cond-mat.mtrl-sci↗

arXiv2015

Molecular simulations of heterogeneous ice nucleation. II. Peeling back the layers

Coarse grained molecular dynamics simulations are presented in which the sensitivity of the ice nucleation rate to the hydrophilicity of a graphene nanoflake is investigated. We find that an optimal interaction strength for promoting ice nucleation exists, which coincides with that found previously for a face centered cubic (111) surface. We further investigate the role that the layering of interfacial water plays in heterogeneous ice nucleation and demonstrate that the extent of layering is not a good indicator of ice nucleating ability for all surfaces. Our results suggest that to be an efficient ice nucleating agent, a surface should not bind water too strongly if it is able to accommodate high coverages of water.

cond-mat.mtrl-sci↗

arXiv2014

On the Accuracy of van der Waals Inclusive Density-Functional Theory Exchange-Correlation Functionals for Ice at Ambient and High Pressures

Density-functional theory (DFT) has been widely used to study water and ice for at least 20 years. However, the reliability of different DFT exchange-correlation (xc) functionals for water remains a matter of considerable debate. This is particularly true in light of the recent development of DFT based methods that account for van der Waals (vdW) dispersion forces. Here, we report a detailed study with several xc functionals (semi-local, hybrid, and vdW inclusive approaches) on ice Ih and six proton ordered phases of ice. Consistent with our previous study [Phys. Rev. Lett. 107, 185701 (2011)] which showed that vdW forces become increasingly important at high pressures, we find here that all vdW inclusive methods considered improve the relative energies and transition pressures of the high-pressure ice phases compared to those obtained with semi-local or hybrid xc functionals. However, we also find that significant discrepancies between experiment and the vdW inclusive approaches remain in the cohesive properties of the various phases, causing certain phases to be absent from the phase diagram. Therefore, room for improvement in the description of water at ambient and high pressures remains and we suggest that because of the stern test the high pressure ice phases pose they should be used in future benchmark studies of simulation methods for water.

cond-mat.mtrl-sci↗