America Forever Bytes
Technology
![[2606.01850] Does Compression Preserve Uncertainty? A Unified Benchmark for Quantized and Sparse LLMs via Conformal Prediction](/static/img/curated/desks/afbytes-tech.png)
Abstract page for arXiv paper 2606.01850: Does Compression Preserve Uncertainty? A Unified Benchmark for Quantized and Sparse LLMs via Conformal Prediction
Abstract page for arXiv paper 2605.30711: SAGE: A Novelty Gate for Efficient Memory Evolution in Agentic LLMs
Abstract page for arXiv paper 2510.14365: Understanding the Ability of LLMs to Handle Character-Level Perturbation
Abstract page for arXiv paper 2602.02103: How Far Ahead Do LLMs Plan? Uncovering the Latent Horizon in Chain-of-Thought Reasoning
Abstract page for arXiv paper 2605.28823: What are They Thinking? Delineation, Probing and Tracking of Concepts in LLMs
Abstract page for arXiv paper 2605.28866: Continuity and Ordinality Matter: Constraining Time Series Tokens for Effective Time Series Analysis with Large Langua...
Abstract page for arXiv paper 2605.28876: LogDx-CI: Benchmarking Log Reduction Tools for LLM Root-Cause Diagnosis
Abstract page for arXiv paper 2605.28889: Context Distillation as Latent Memory Management
Abstract page for arXiv paper 2605.30014: From GPS Points to Travel Patterns: Flexible and Semantic Trajectory Generation with LLMs
Abstract page for arXiv paper 2605.30029: RAISE: RAG Design as an Architecture Search Problem
Abstract page for arXiv paper 2605.30036: Teaching Values to Machines: Simulating Human-Like Behavior in LLMs
Abstract page for arXiv paper 2605.30039: Domain-Specific Data Synthesis for LLMs via Minimal Sufficient Representation Learning
Abstract page for arXiv paper 2605.30042: Learning to Choose: An Empowerment-Guided Multi-Agent System with semantic communication for Adaptive Method Selection
Abstract page for arXiv paper 2605.28277: Do LLMs Build World Models From Text? A Multilingual Diagnostic of Spatial Reasoning
Abstract page for arXiv paper 2605.28591: Models That Know How Evaluations Are Designed Score Safer
Abstract page for arXiv paper 2605.28598: Evaluating the Realism of LLM-powered Social Agents: A Case Study of Reactions to Spanish Online News
Abstract page for arXiv paper 2605.28600: Transformers Provably Learn to Internalize Chain-of-Thought