Benchmarking at the Edge of Comprehension

As frontier Large Language Models (LLMs) increasingly saturate new benchmarks shortly after they are published, benchmarking itself is at a juncture: if frontier models keep improving, it will become increasingly hard for humans to generate discriminative tasks, provide accurate ground-truth answers, or evaluate complex solutions. If benchmarking becomes infeasible, our ability to measure any progress in AI is at stake. We refer to this scenario as the post-comprehension regime. In this work, we propose Critique-Resilient Benchmarking, an adversarial framework designed to compare models even when full human understanding is infeasible. Our technique relies on the notion of critique-resilient correctness: an answer is deemed correct if no adversary has convincingly proved otherwise. Unlike standard benchmarking, humans serve as bounded verifiers and focus on localized claims, which preserves evaluation integrity beyond full comprehension of the task. Using an itemized bipartite Bradley-Terry model, we jointly rank LLMs by their ability to solve challenging tasks and to generate difficult yet solvable questions. We showcase the effectiveness of our method in the mathematical domain across eight frontier LLMs, showing that the resulting scores are stable and correlate with external capability measures. Our framework reformulates benchmarking as an adversarial generation-evaluation game in which humans serve as final adjudicators.

Tacit Coordination of Large Language Models

In tacit coordination games with multiple outcomes, purely rational solution concepts, such as Nash equilibria, provide no guidance for which equilibrium to choose. Shelling's theory explains how, in these settings, humans coordinate by relying on focal points: solutions or outcomes that naturally arise because they stand out in some way as salient or prominent to all players. This work studies Large Language Models (LLMs) as players in tacit coordination games, and addresses how, when, and why focal points emerge. We compare and quantify the coordination capabilities of LLMs in cooperative and competitive games for which human experiments are available. We also introduce several learning-free strategies to improve the coordination of LLMs, with themselves and with humans. On a selection of heterogeneous open-source models, including Llama, Qwen, and GPT-oss, we discover that LLMs have a remarkable capability to coordinate and often outperform humans, yet fail on common-sense coordination that involves numbers or nuanced cultural archetypes. This paper constitutes the first large-scale assessment of LLMs' tacit coordination within the theoretical and psychological framework of focal points.

An End-to-end Planning Framework with Agentic LLMs and PDDL

We present an end-to-end framework for planning supported by verifiers. An orchestrator receives a human specification written in natural language and converts it into a PDDL (Planning Domain Definition Language) model, where the domain and problem are iteratively refined by sub-modules (agents) to address common planning requirements, such as time constraints and optimality, as well as ambiguities and contradictions that may exist in the human specification. The validated domain and problem are then passed to an external planning engine to generate a plan. The orchestrator and agents are powered by Large Language Models (LLMs) and require no human intervention at any stage of the process. Finally, a module translates the final plan back into natural language to improve human readability while maintaining the correctness of each step. We demonstrate the flexibility and effectiveness of our framework across various domains and tasks, including the Google NaturalPlan benchmark and PlanBench, as well as planning problems like Blocksworld and the Tower of Hanoi (where LLMs are known to struggle even with small instances). Our framework can be integrated with any PDDL planning engine and validator (such as Fast Downward, LPG, POPF, VAL, and uVAL, which we have tested) and represents a significant step toward end-to-end planning aided by LLMs.

Large Language Models Miss the Multi-Agent Mark

Recent interest in Multi-Agent Systems of Large Language Models (MAS LLMs) has led to an increase in frameworks leveraging multiple LLMs to tackle complex tasks. However, much of this literature appropriates the terminology of MAS without engaging with its foundational principles. In this position paper, we highlight critical discrepancies between MAS theory and current MAS LLMs implementations, focusing on four key areas: the social aspect of agency, environment design, coordination and communication protocols, and measuring emergent behaviours. Our position is that many MAS LLMs lack multi-agent characteristics such as autonomy, social interaction, and structured environments, and often rely on oversimplified, LLM-centric architectures. The field may slow down and lose traction by revisiting problems the MAS literature has already addressed. Therefore, we systematically analyse this issue and outline associated research opportunities; we advocate for better integrating established MAS concepts and more precise terminology to avoid mischaracterisation and missed opportunities.

Fixed Point Explainability

This paper introduces a formal notion of fixed point explanations, inspired by the "why regress" principle, to assess, through recursive applications, the stability of the interplay between a model and its explainer. Fixed point explanations satisfy properties like minimality, stability, and faithfulness, revealing hidden model behaviours and explanatory weaknesses. We define convergence conditions for several classes of explainers, from feature-based to mechanistic tools like Sparse AutoEncoders, and we report quantitative and qualitative results.

Language Models Are Implicitly Continuous

Language is typically modelled with discrete sequences. However, the most successful approaches to language modelling, namely neural networks, are continuous and smooth function approximators. In this work, we show that Transformer-based language models implicitly learn to represent sentences as continuous-time functions defined over a continuous input space. This phenomenon occurs in most state-of-the-art Large Language Models (LLMs), including Llama2, Llama3, Phi3, Gemma, Gemma2, and Mistral, and suggests that LLMs reason about language in ways that fundamentally differ from humans. Our work formally extends Transformers to capture the nuances of time and space continuity in both input and output space. Our results challenge the traditional interpretation of how LLMs understand language, with several linguistic and engineering implications.

Understanding the Logical Capabilities of Large Language Models via Out-of-Context Representation Learning

We study the capabilities of Large Language Models (LLM) on binary relations, a ubiquitous concept in math employed in most reasoning, math and logic benchmarks. This work focuses on equality, inequality, and inclusion, along with the properties they satisfy, such as ir/reflexivity, a/symmetry, transitivity, and logical complexity (e.g., number of reasoning ``hops''). We propose an alternative to in-context learning that trains only the representations of newly introduced tokens, namely out-of-context representation learning. This method mitigates linguistic biases already present in a model and, differently from in-context learning, does not rely on external information or illustrations. We argue out-of-context representation learning as a better alternative to in-context learning and fine-tuning to evaluate the capabilities of LLMs on logic tasks that are the building blocks of more complex reasoning benchmarks.

When Claims Evolve: Evaluating and Enhancing the Robustness of Embedding Models Against Misinformation Edits

Online misinformation remains a critical challenge, and fact-checkers increasingly rely on embedding-based methods to retrieve relevant fact-checks. Yet, when debunked claims reappear in edited forms, the performance of these methods is unclear. In this work, we introduce a taxonomy of six common real-world misinformation edits and propose a perturbation framework that generates valid, natural claim variations. Our multi-stage retrieval evaluation reveals that standard embedding models struggle with user-introduced edits, while LLM-distilled embeddings offer improved robustness at a higher computational cost. Although a strong reranker helps mitigate some issues, it cannot fully compensate for first-stage retrieval gaps. Addressing these retrieval gaps, our train- and inference-time mitigation approaches enhance in-domain robustness by up to 17 percentage points and boost out-of-domain generalization by 10 percentage points over baseline models. Overall, our findings provide practical improvements to claim-matching systems, enabling more reliable fact-checking of evolving misinformation.

Code Simulation as a Proxy for High-order Tasks in Large Language Models

Many reasoning, planning, and problem-solving tasks share an intrinsic algorithmic nature: correctly simulating each step is a sufficient condition to solve them correctly. We collect pairs of naturalistic and synthetic reasoning tasks to assess the capabilities of Large Language Models (LLM). While naturalistic tasks often require careful human handcrafting, we show that synthetic data is, in many cases, a good proxy that is much easier to collect at scale. We leverage common constructs in programming as the counterpart of the building blocks of naturalistic reasoning tasks, such as straight-line programs, code that contains critical paths, and approximate and redundant instructions. We further assess the capabilities of LLMs on sorting problems and repeated operations via sorting algorithms and nested loops. Our synthetic datasets further reveal that while the most powerful LLMs exhibit relatively strong execution capabilities, the process is fragile: it is negatively affected by memorisation and seems to rely heavily on pattern recognition. Our contribution builds upon synthetically testing the reasoning capabilities of LLMs as a scalable complement to handcrafted human-annotated problems.

Jailbreaking Large Language Models in Infinitely Many Ways

We discuss the "Infinitely Many Meanings" attacks (IMM), a category of jailbreaks that leverages the increasing capabilities of a model to handle paraphrases and encoded communications to bypass their defensive mechanisms. IMMs' viability pairs and grows with a model's capabilities to handle and bind the semantics of simple mappings between tokens and work extremely well in practice, posing a concrete threat to the users of the most powerful LLMs in commerce. We show how one can bypass the safeguards of the most powerful open- and closed-source LLMs and generate content that explicitly violates their safety policies. One can protect against IMMs by improving the guardrails and making them scale with the LLMs' capabilities. For two categories of attacks that are straightforward to implement, i.e., bijection and encoding, we discuss two defensive strategies, one in token and the other in embedding space. We conclude with some research questions we believe should be prioritised to enhance the defensive mechanisms of LLMs and our understanding of their safety.