Neural networks excel across a wide range of tasks, yet remain black boxes. In particular, how their internal representations are shaped by the complexity of the input data and the problems they solve remains obscure. In this work, we introduce a suite of five data-agnostic probes-pruning, binarization, noise injection, sign flipping, and bipartite network randomization-to quantify how task difficulty influences the topology and robustness of representations in multilayer perceptrons (MLPs). MLPs are represented as signed, weighted bipartite graphs from a network science perspective. We contrast easy and hard classification tasks on the MNIST and Fashion-MNIST datasets. We show that binarizing weights in hard-task models collapses accuracy to chance, whereas easy-task models remain robust. We also find that pruning low-magnitude edges in binarized hard-task models reveals a sharp phase-transition in performance. Moreover, moderate noise injection can enhance accuracy, resembling a stochastic-resonance effect linked to optimal sign flips of small-magnitude weights. Finally, preserving only the sign structure-instead of precise weight magnitudes-through bipartite network randomizations suffices to maintain high accuracy. These phenomena define a model- and modality-agnostic measure of task complexity: the performance gap between full-precision and binarized or shuffled neural network performance. Our findings highlight the crucial role of signed bipartite topology in learned representations and suggest practical strategies for model compression and interpretability that align with task complexity.

This paper introduces and formalizes Noosemìa, a novel cognitive-phenomenological pattern emerging from human interaction with generative AI systems, particularly those enabling dialogic or multimodal exchanges. We propose a multidisciplinary framework to explain how, under certain conditions, users attribute intentionality, agency, and even interiority to these systems - a process grounded not in physical resemblance, but in linguistic performance, epistemic opacity, and emergent technological complexity. By linking an LLM declination of meaning holism to our technical notion of the LLM Contextual Cognitive Field, we clarify how LLMs construct meaning relationally and how coherence and a simulacrum of agency arise at the human-AI interface. The analysis situates noosemia alongside pareidolia, animism, the intentional stance and the uncanny valley, distinguishing its unique characteristics. We also introduce a-noosemia to describe the phenomenological withdrawal of such projections. The paper concludes with reflections on the broader philosophical, epistemological and social implications of noosemic dynamics and directions for future research.

Incorporating external knowledge is crucial for knowledge-intensive tasks, such as question answering and fact checking. However, language models (LMs) may ignore relevant information that contradicts outdated parametric memory or be distracted by irrelevant contexts. While many context utilisation manipulation techniques (CMTs) have recently been proposed to alleviate these issues, few have seen systematic comparison. In this paper, we develop CUB (Context Utilisation Benchmark) - the first comprehensive benchmark designed to help practitioners within retrieval-augmented generation (RAG) diagnose CMTs under different context conditions. With this benchmark, we conduct the most extensive evaluation to date of seven state-of-the-art methods, representative of the main categories of CMTs, across three diverse datasets and tasks, applied to nine LMs. Our results reveal that most existing CMTs struggle to handle the full spectrum of context types encountered in real-world retrieval-augmented scenarios. We also find that many CMTs display inflated performance on simple synthesised datasets, compared to more realistic datasets with naturally occurring samples. Our findings expose critical gaps in current CMT evaluation practices and demonstrate the need for holistic testing and the development of CMTs that can robustly handle multiple context types.

How do the latent spaces used by independently-trained LLMs relate to one another? We study the nearest neighbor relationships induced by activations at different layers of 24 open-weight LLMs, and find that they 1) tend to vary from layer to layer within a model, and 2) are approximately shared between corresponding layers of different models. Claim 2 shows that these nearest neighbor relationships are not arbitrary, as they are shared across models, but Claim 1 shows that they are not "obvious" either, as there is no single set of nearest neighbor relationships that is universally shared. Together, these suggest that LLMs generate a progression of activation geometries from layer to layer, but that this entire progression is largely shared between models, stretched and squeezed to fit into different architectures.

Palantir's meteoric rise is pushing the company's valuation further into record territory, forcing bullish investors to bank on increasingly robust future growth to justify its current level. Shares of the defense maker closed at another all-time high Friday, bringing gains since its 2021 debut to near 2,500%. The stock is up almost 150% this year, a rally underpinned by the company's growing use of artificial intelligence, business ties to the U.S. government and most recently, a stellar earnings report. That surge has made Palantir eye-wateringly expensive compared to its peers: trading at 245 times forward earnings, it is the most richly-valued company in the S&P 500 Index.

Such models, recently coined foundation models, have been transformational to the field of natural language processing.

Ukraine's military has said it struck an oil refinery in Russia's Saratov region in an overnight drone attack, causing explosions and destruction, according to an army statement, as daily aerial exchanges intensify with diplomatic momentum to end the war in play. Saratov's governor said on Sunday that one person was killed and several residential apartments and an industrial facility were damaged, but did not mention the oil refinery being struck. "[Ukrainian] drones are targeting … deeper into Russian territory [than] in the past, where previous attacks have been focused on the line of contact in the south and the western parts of Russia," said Al Jazeera's Osama Bin Javaid, reporting from Moscow. It is still unclear whether Ukraine's claims that it hit a refinery are true, he added. Ukraine's military also said on Sunday that it had taken back a village in the Sumy region from the Russian army, which has made significant recent gains there.

The spark becomes a flame, and within seconds, a satellite dish swirling overhead picks up on the anomaly and triggers an alarm. An autonomous helicopter takes flight and zooms toward the fire, using sensors to locate the blaze and artificial intelligence to generate a plan of attack. It measures the wind speed and fire movement, communicating constantly with the unmanned helicopter behind it, and the one behind that. Los Angeles knows how to weather a crisis -- or two or three. Angelenos are tapping into that resilience, striving to build a city for everyone.

Staff at the UK's leading artificial intelligence institute have raised concerns about the organisation's governance and internal culture in a whistleblowing complaint to the charity watchdog. The Alan Turing Institute (ATI), a registered charity with substantial state funding, is under government pressure to overhaul its strategic focus and leadership after an intervention last month from the technology secretary, Peter Kyle. In a complaint to the Charity Commission, a group of current ATI staff raise eight points of concern and say the institute is in danger of collapse due to government threats over its funding. The complaint alleges that the board of trustees, chaired by the former Amazon UK boss Doug Gurr, has failed to fulfil core legal duties such as providing strategic direction and ensuring accountability, with staff alleging a letter of no confidence was delivered last year and not acted upon. A spokesperson for ATI said the Charity Commission had not been in touch with the institute about any complaints that may have been sent to the organisation.

Russian forces launched a drone attack on a bus in Ukraine's Kherson region, killing at least two people and wounding 16 others, according to Ukrainian officials. Another drone hit the bus as the police were responding to the attack, injuring three officers, the police added. Russian forces also launched 36 other attacks on settlements across the Kherson region through Friday and Saturday, killing at least one more person and injuring three, according to Governor Oleksandr Prokudin. Russian attacks on Ukraine's Zaporizhia region killed two people travelling in a car in the Bilenkivska community on Saturday morning, and a 61-year-old woman who was in her home in the Vasylivka district, a local official reported. In Ukraine's Dnipropetrovsk region, a Russian attack killed a 56-year-old woman and wounded a 62-year-old man in the city of Nikopol, while in the Donetsk region, other Russian attacks killed four people and wounded nine, according to officials.