Find a story
Search Spins
Search titles, summaries, and missing voices across published articles — press releases, announcements, and media coverage.
12 results for “embeddings”
Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis
Hugging Face announced OlmoEarth embeddings, a new feature allowing users to export custom embeddings from its OlmoEarth Studio platform for downstream analysis.
Aug 12, 2026
Position Encoding in Transformers: From Absolute and Relative Methods to Rotary Position Embeddings and Long-Context Scaling
A technical survey paper on position encoding methods in Transformers synthesizes and compares absolute, relative, and rotary embedding techniques, with emphasis on long-context scaling strategies and empirical evaluation criteria.
Aug 13, 2026
A Study of ASR Adaptation and Representation Dimensionality Reduction in Persian Speech Emotion Recognition Using Whisper
Researchers adapted Whisper for Persian Speech Emotion Recognition (SER) using PCA-based dimensionality reduction to cut parameters and training costs, finding it improves performance on the ShEMO dataset while ASR fine-tuning delivered only modest SER gains.
Aug 7, 2026
SJEPA: Learning Elegant Latent Dynamics with Hybrid Symbolic-Neural Predictors
SJEPA is a new joint-embedding predictive architecture that integrates symbolic rules with neural corrections to learn interpretable, low-complexity latent dynamics — advancing the goal of making AI models' internal state transitions both predictive and human-understandable.
Aug 6, 2026
Guarantees on Dynamical System Distinguishability for LLM Token Generation
A theoretical paper establishes formal guarantees for distinguishing LLM-generated text by modeling token embeddings as stochastic linear dynamical systems and proving exponential decay in misclassification probability with sequence length.
Aug 3, 2026
FloDR: An invertible dimensionality reduction method based on a normalising flow
FloDR is a new invertible dimensionality reduction method that preserves unused dimensions to enable diagnostic visualizations—like conditional spread and hidden contrast—with statistical confidence testing, addressing interpretability limits of t-SNE and UMAP.
Jul 30, 2026
Hierarchical Grading in Large Language Models
Researchers propose Graded Large Language Models (GLLMs), a theoretical extension of transformer architecture using algebraic grading to improve statistical efficiency for level-stratified prediction tasks, with claims of provable risk separation and pre-certified optimization.
Jul 28, 2026
emb-diversity: A Tool for Embedding-Based Measurement of Data Diversity
A new open-source tool called emb-diversity provides standardized, embedding-based methods to measure data diversity across stylistic, semantic, language, and speaker dimensions — addressing a fragmentation in NLP evaluation practices.
Jul 23, 2026
14× faster embeddings: how we rebuilt the ONNX path in Manticore
A community discussion on Hacker News about performance improvements to the ONNX inference path in Manticore, an open-source AI model serving framework, claiming 14× faster embeddings generation.
Published Jul 3, 2026 · Analyzed Jul 6, 2026
DiscoLoop: Looping Discrete Embeddings and Continuous Hidden States for Multi-hop Reasoning
Researchers propose a new architecture for multi-hop reasoning tasks in large language models.
Published Jul 2, 2026 · Analyzed Jul 5, 2026
ALEE: Any-Language Evaluation of Embeddings via English-Centric Minimal Pairs
Researchers introduced ALEE, a new cross-lingual evaluation framework for text embeddings that uses English-centric minimal pairs grounded in Abstract Meaning Representations to assess semantic fidelity across 275+ languages — addressing longstanding limitations in static, narrow, and overfit embedding benchmarks.
Published Jul 2, 2026 · Analyzed Jul 5, 2026
Inside Target’s LLM-Based System for Semantic Matching in Marketing Forecast Pipelines
Target deployed an internal LLM-based semantic matching system to automate and improve marketing campaign forecasting by retrieving and ranking analogous past campaigns, replacing manual, rule-based processes.
Published Jun 29, 2026 · Analyzed Jul 4, 2026