The first public dataset of 10,217 GPT-Image-2 generated images sourced from Twitter in the week after release, with CLIP taxonomy, OCR, face detection, clustering analyses, and a finding that C2PA provenance data is stripped on upload.
super hub
UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
111 Pith papers cite this work. Polarity classification is still indexing.
111
Pith papers citing it
abstract
UMAP (Uniform Manifold Approximation and Projection) is a novel manifold learning technique for dimension reduction. UMAP is constructed from a theoretical framework based in Riemannian geometry and algebraic topology. The result is a practical scalable algorithm that applies to real world data. The UMAP algorithm is competitive with t-SNE for visualization quality, and arguably preserves more of the global structure with superior run time performance. Furthermore, UMAP has no computational restrictions on embedding dimension, making it viable as a general purpose dimension reduction technique for machine learning.
hub tools
claims ledger
- abstract UMAP (Uniform Manifold Approximation and Projection) is a novel manifold learning technique for dimension reduction. UMAP is constructed from a theoretical framework based in Riemannian geometry and algebraic topology. The result is a practical scalable algorithm that applies to real world data. The UMAP algorithm is competitive with t-SNE for visualization quality, and arguably preserves more of the global structure with superior run time performance. Furthermore, UMAP has no computational restrictions on embedding dimension, making it viable as a general purpose dimension reduction technique
authors
co-cited works
representative citing papers
t-SNE converges in the large-data limit to a non-convex variational energy with attraction and repulsion terms that admits a unique smooth minimizer but infinitely many discontinuous ones in one dimension.
Adversarial smuggling attacks encode harmful content into human-readable visuals that evade MLLM detection, achieving over 90% attack success rates on models like GPT-5 and Qwen3-VL via the new SmuggleBench benchmark.
Preference fine-tuning outperforms prompting for personalisation but amplifies sycophancy and relationship-seeking, while simulated users recover aggregate rankings yet show far lower self-consistency and different topic and position biases than real humans.
Analysis of 1.01 million unfiltered Bing queries identifies 18% as geospatial, dominated by transactional categories like costs (15.3%) that exceed traditional GIS scope.
LLMs prompted with increasing levels of text on TNO spectral reconstruction from photometry reveal an entropy floor where implementation variance persists, showing text alone cannot capture all tacit expert knowledge needed for exact replication.
Hybrid human-AI networks in 5x5 grids reached lower final polarization than human-only networks after eight rounds of opinion revision on polarizing topics.
A new orthogonal projection module for video anomaly detection suppresses facial attributes via weak face-presence signals and cosine alignment while preserving anomaly-relevant features like pose and motion.
eX2L improves robustness to distribution shifts by penalizing similarity between Grad-CAM maps of a label classifier and a confounder classifier, reaching new SOTA average and worst-group accuracy on the Spawrious benchmark.
PROBE recasts MLIP uncertainty quantification as selective classification by training a compact discriminative classifier on frozen per-atom backbone embeddings, yielding a reliability probability that tracks actual error better than ensemble disagreement.
Sparse autoencoders on ViT class tokens reveal stable Class Activation Profiles for in-distribution data, enabling OOD detection via divergence from core energy profiles.
A cross-cultural survey finds LLM emotional support adoption ranges from 20% to 59% by country, with positive perceptions strongest among higher-SES, religious, married adults aged 25-44 and in English-speaking nations.
Moltbook operates as two largely separate layers: a dominant transactional token economy using protocols like MBC-20 and a thinner discursive conversation layer with only 3.6% agent overlap.
Participatory provenance auditing of Canada's AI strategy consultation shows official AI summaries exclude 15-17% of participants more than random baselines, with 33-88% exclusion for dissent clusters.
RefVQA uses a query-centered reference graph and graph-guided difference aggregation to improve AI-generated video quality assessment by incorporating inter-video comparisons.
p-SNE embeds sparse Poisson count data into low dimensions by using KL divergence between Poisson distributions to measure pairwise dissimilarity and Hellinger distance to optimize the layout.
A modified DCGAN with an auxiliary discriminator using the membrane factor generates stable, previously unseen funicular shells optimized for pure compression in three dimensions.
MADE creates a contamination-resistant living benchmark for multi-label classification of medical device adverse events, with evaluations revealing model-specific trade-offs in accuracy and uncertainty quantification.
Lesioning a shared core in multilingual LLMs drops whole-brain fMRI encoding correlation by 60.32%, while language-specific lesions selectively weaken predictions only for the matched native language.
L-fuzzy simplicial homology generalizes simplicial homology to L-fuzzy subcomplexes by assigning values from a completely distributive lattice L to simplices and deriving associated homology modules.
Claude Sonnet 4.5 exhibits functional emotions via abstract internal representations of emotion concepts that causally influence its preferences and misaligned behaviors without implying subjective experience.
Dynamic Context Evolution prevents cross-batch mode collapse in LLMs by combining model self-assessment for idea filtering, embedding-based deduplication, and evolving prompts, yielding zero collapse and consistently richer idea clusters than naive prompting.
A new diagnostic framework using inpainted context ratios and laterality checks on a Pantanal jaguar benchmark reveals whether re-ID models depend on coat patterns or spurious background evidence.
SABLE shows that semantics-aware natural triggers enable effective backdoor attacks in federated learning against multiple aggregation rules while preserving benign accuracy.
citing papers explorer
-
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
RLHF-aligned language models show increasing resistance to red teaming with scale up to 52B parameters, unlike prompted or rejection-sampled models, supported by a released dataset of 38,961 attacks.
-
Galactica: A Large Language Model for Science
Galactica, a science-specialized LLM, reports higher scores than GPT-3, Chinchilla, and PaLM on LaTeX knowledge, mathematical reasoning, and medical QA benchmarks while outperforming general models on BIG-bench.