Tag
HyperSAE is an open-source Python library that uses hyperbolic geometry to organize LLM learned concepts into browsable tree structures, improving on flat feature lists. It captures 99.8% of Gemma-2-2B's features and includes interactive demos.
Introduces tvscreener, a Python library that can directly fetch TradingView stock screener data, supports multiple markets and 13,000+ technical indicators, returns Pandas DataFrame, and is convenient for quantitative strategy use.
A simple Python library retrieves data from TradingView Screener, featuring a visual code generator for 13,000+ fields and Model Context Protocol (MCP) support for AI assistants.
circuit-tracer is a Python library that helps interpretability researchers map model computations by calculating direct effects between transcoder features and output logits, then visualizing them as an interactive attribution graph.
The article revisits the gpiozero library's source/values abstraction for connecting GPIO devices declaratively, and proposes a drag-and-drop web UI that would let users visually wire devices together, building on the library's flow-oriented design.
XY is a fast, composable GPU-accelerated interactive plotting library for Python, designed for web and notebooks, handling up to 10 billion points with pan and zoom.
Introduces grapheme-kit, an open-source Python library that extends lexical distance and evaluation metrics to operate on grapheme clusters instead of Unicode code points, with improved processing for Tamil and Sinhala.
GigaToken is an ultra-fast tokenizer library that claims up to 1000x speedup over HuggingFace tokenizers, supporting most common LLM tokenizers and providing drop-in compatibility.
Goldman Sachs open-sourced its internal quantitative trading toolkit gs-quant, providing institutional-grade derivatives pricing, risk management, and strategy development tools. It has received 11.3K GitHub stars.
token-budget-contracts is a library for governing token spend across multi-agent systems. Version 0.3.0 adds native adapters for LangGraph and CrewAI, plus OpenTelemetry integration for monitoring budget activity.
AuditWeave is a lightweight Python library that records workflow steps into a tamper-evident, hash-chained ledger for auditing AI-assisted and data-transformation workflows, enabling traceability and integrity verification.
SPEAR is a Python library that controls Unreal Engine for photorealistic rendering at high speed, providing extensive programmability and unique ground-truth modalities for embodied AI research.
Object Aligner is an open-source Python library that deterministically scores two JSON objects by recursively aligning their trees, using Hungarian algorithm for unordered collections and sequence alignment for ordered ones. It introduces referential alignment for graphs/hypergraphs and can be used as a reward function in LLM prompt optimization.
PyMuPDF 1.28 adds first-class Markdown support, allowing PDF creation from Markdown text with CSS control.
Microsoft open-sourced SkillOpt, a framework that treats agent skill documents as trainable artifacts, optimizing them through iterative feedback loops without modifying model weights. It is available as a Python library with support for self-evolving agent skills.
EverMind team under Shanda Group founder Chen Tianqiao has open-sourced EverMemOS (EverOS), a brain-inspired intelligent memory operating system designed to solve the forgetting and context-limited problems of large language models by simulating human memory mechanisms, providing long-term memory infrastructure for AI Agents.
This article introduces nomograms as graphical calculators and describes the pyNomo library for creating them programmatically.
DesignArena open-sources a Python library that converts HTML/CSS to editable PowerPoint files using headless Chromium, sharing lessons from generating over 50,000 AI slides.
RAMPART is a Python library that makes LLM context assembly programmable, allowing developers to register named blocks of context for placement before the model's first token. It improves performance by tens of percentage points on various models through block clustering and tool access control.
PyCC.id is a Python library for hypothesis-driven equation discovery from time-series data, leveraging structural identifiability to help filter candidate models.