system2-attention
System 2 Attention Skill: Deliberate Reasoning Validation
Status: ✅ Production Ready Trit: -1 (MINUS - validator/constraint) Color: #2626D8 (Blue) Principle: Filter noise via deliberate re-attention Frame: Two-stage attention with explicit reasoning
Overview
System 2 Attention (S2A) validates and filters transformer attention by regenerating context deliberately. Standard attention (System 1) is fast but susceptible to sycophancy and irrelevant context. S2A re-attends after explicit reasoning.
- Context regeneration: LLM rewrites context removing irrelevant info
- Two-pass attention: Fast then deliberate
- Sycophancy reduction: Filter opinion-seeking noise
- Factual grounding: Anchor to verified facts
More from plurigrid/asi
academic-research
Search academic papers across arXiv, PubMed, Semantic Scholar, bioRxiv, medRxiv, Google Scholar, and more. Get BibTeX citations, download PDFs, analyze citation networks. Use for literature reviews, finding papers, and academic research.
59wev-tesseract
WEV Tesseract Skill
33tree-sitter
AST-based code analysis using tree-sitter. Use for parsing code structure, extracting symbols, finding patterns with tree-sitter queries, analyzing complexity, and understanding code architecture. Supports Python, JavaScript, TypeScript, Go, Rust, C, C++, Swift, Java, Kotlin, Julia, and more.
22reverse-engineering
Reverse Engineering Skill
17bdd-mathematical-verification
BDD-Driven Mathematical Content Verification Skill
17alife
Comprehensive Artificial Life skill combining ALIFE2025 proceedings, classic texts (Axelrod, Epstein-Axtell), ALIEN simulation, Lenia, NCA, swarm intelligence, and evolutionary computation. 337 pages extracted, 80+ papers, 153 figures.
16