Tootfinder

This makes me a bit nervous as I really love this game. However some speculation I’ve seen is the Sasquatch could become a much-needed mascot for Apple Arcade.
https://www.macobserver.com/news/apple-acquires-indie-video-game-studio-rac7/

Last week, we continued our #ISE2025 lecture on distributional semantics with the introduction of neural language models (NLMs) and compared them to traditional statistical n-gram models.
Benefits of NLMs:
- Capturing Long-Range Dependencies
- Computational and Statistical Tractability
- Improved Generalisation
- Higher Accuracy
@…

This https://arxiv.org/abs/2502.19679 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csDL_…

Architectural Vulnerability and Reliability Challenges in AI Text Annotation: A Survey-Inspired Framework with Independent Probability Assessment
Large Language Models, despite their power, have a fundamental architectural vulnerability stemming from their causal transformer design -- order sensitivity. This architectural constraint may distorts classification outcomes when prompt elements like label options are reordered, revealing a theoretical gap between accuracy metrics and true model reliability. The paper conceptualizes this vulnerability through the lens of survey methodology, where respondent biases parallel LLM positional depen…

Resonance Complexity Theory and the Architecture of Consciousness: A Field-Theoretic Model of Resonant Interference and Emergent Awareness
Michael Arnold Bruna
https://arxiv.org/abs/2505.20580

Resonance Complexity Theory and the Architecture of Consciousness: A Field-Theoretic Model of Resonant Interference and Emergent Awareness
This paper introduces Resonance Complexity Theory (RCT), which proposes that consciousness emerges from stable interference patterns of oscillatory neural activity. These patterns, shaped by recursive feedback and constructive interference, must exceed critical thresholds in complexity, coherence, gain, and fractal dimensionality to give rise to conscious experience. The resulting spatiotemporal attractors encode subjective awareness as dynamic resonance structures distributed across the neural…

TopK Language Models
Ryosuke Takahashi, Tatsuro Inaba, Kentaro Inui, Benjamin Heinzerling
https://arxiv.org/abs/2506.21468 https://arxiv.org/pdf/2506.21468 https://arxiv.org/html/2506.21468
arXiv:2506.21468v1 Announce Type: new
Abstract: Sparse autoencoders (SAEs) have become an important tool for analyzing and interpreting the activation space of transformer-based language models (LMs). However, SAEs suffer several shortcomings that diminish their utility and internal validity. Since SAEs are trained post-hoc, it is unclear if the failure to discover a particular concept is a failure on the SAE's side or due to the underlying LM not representing this concept. This problem is exacerbated by training conditions and architecture choices affecting which features an SAE learns. When tracing how LMs learn concepts during training, the lack of feature stability also makes it difficult to compare SAEs features across different checkpoints. To address these limitations, we introduce a modification to the transformer architecture that incorporates a TopK activation function at chosen layers, making the model's hidden states equivalent to the latent features of a TopK SAE. This approach eliminates the need for post-hoc training while providing interpretability comparable to SAEs. The resulting TopK LMs offer a favorable trade-off between model size, computational efficiency, and interpretability. Despite this simple architectural change, TopK LMs maintain their original capabilities while providing robust interpretability benefits. Our experiments demonstrate that the sparse representations learned by TopK LMs enable successful steering through targeted neuron interventions and facilitate detailed analysis of neuron formation processes across checkpoints and layers. These features make TopK LMs stable and reliable tools for understanding how language models learn and represent concepts, which we believe will significantly advance future research on model interpretability and controllability.
toXiv_bot_toot

Refining Datapath for Microscaling ViTs
Can Xiao, Jianyi Cheng, Aaron Zhao
https://arxiv.org/abs/2505.22194 https://arxiv.org/pdf/250…

Refining Datapath for Microscaling ViTs
Vision Transformers (ViTs) leverage the transformer architecture to effectively capture global context, demonstrating strong performance in computer vision tasks. A major challenge in ViT hardware acceleration is that the model family contains complex arithmetic operations that are sensitive to model accuracy, such as the Softmax and LayerNorm operations, which cannot be mapped onto efficient hardware with low precision. Existing methods only exploit parallelism in the matrix multiplication ope…

Some wear balaclavas. Some wearneck gators, sunglasses and hats. Some wear masks and casual clothes.
Across the country, armed federal immigration officers have increasingly hidden their identities while carrying out immigration raids, arresting protesters and roughing upprominent Democratic critics.
It’s a trend that has sparked alarm among civil rights and law enforcement experts alike.
Mike German, a former FBIagent, said officers’ widespread use of masks was unprecedent…

Adaptive Hybrid Sort: Dynamic Strategy Selection for Optimal Sorting Across Diverse Data Distributions
Shrinivass Arunachalam Balasubramanian
https://arxiv.org/abs/2506.20677

Adaptive Hybrid Sort: Dynamic Strategy Selection for Optimal Sorting Across Diverse Data Distributions
Sorting is an essential operation in computer science with direct consequences on the performance of large scale data systems, real-time systems, and embedded computation. However, no sorting algorithm is optimal under all distributions of data. The new adaptive hybrid sorting paradigm proposed in this paper is the paradigm that automatically selects the most effective sorting algorithm Counting Sort, Radix Sort, or QuickSort based on real-time monitoring of patterns in input data. The architec…

WAFT: Warping-Alone Field Transforms for Optical Flow
Yihan Wang, Jia Deng
https://arxiv.org/abs/2506.21526 https://arxiv.org/pdf/250…

WAFT: Warping-Alone Field Transforms for Optical Flow
We introduce Warping-Alone Field Transforms (WAFT), a simple and effective method for optical flow. WAFT is similar to RAFT but replaces cost volume with high-resolution warping, achieving better accuracy with lower memory cost. This design challenges the conventional wisdom that constructing cost volumes is necessary for strong performance. WAFT is a simple and flexible meta-architecture with minimal inductive biases and reliance on custom designs. Compared with existing methods, WAFT ranks 1s…

This https://arxiv.org/abs/2505.16968 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csAR_…

CASS: Nvidia to AMD Transpilation with Data, Models, and Benchmark
We introduce \texttt{CASS}, the first large-scale dataset and model suite for cross-architecture GPU code transpilation, targeting both source-level (CUDA~$\leftrightarrow$~HIP) and assembly-level (Nvidia SASS~$\leftrightarrow$~AMD RDNA3) translation. The dataset comprises 70k verified code pairs across host and device, addressing a critical gap in low-level GPU code portability. Leveraging this resource, we train the \texttt{CASS} family of domain-specific language models, achieving 95\% sourc…