Thinking on the Fly: Test-Time Reasoning Enhancement via Latent Thought Policy Optimization
arxiv.org·3h
💻Local LLMs
Toy Binary Decision Diagrams
philipzucker.com·1d
🧮Algebraic Datatypes
LLM Optimization Notes: Memory, Compute and Inference Techniques
gaurigupta19.github.io·15h·
Discuss: Hacker News
💻Local LLMs
Cactus Language • Semantics 1
inquiryintoinquiry.com·15h
🔢Denotational Semantics
Prompting Techniques for Specialised LLMs
dev.to·1d·
Discuss: DEV
🔗Constraint Handling
Certifiable Safe RLHF: Fixed-Penalty Constraint Optimization for Safer Language Models
arxiv.org·3h
🔗Parser Combinators
Atomic and Saturated Models
functor.network·3d·
Discuss: Hacker News
🔢Denotational Semantics
Causal Abstractions, Categorically Unified
arxiv.org·3h
✨Effect Handlers
PLSEMANTICSBENCH: Large Language Models As Programming Language Interpreters
arxiv.org·3h
💻Programming languages
Harnessing LLM for Noise-Robust Cognitive Diagnosis in Web-Based Intelligent Education Systems
arxiv.org·3h
🧠Intelligence Compression
Small Language Models for Agentic Systems: A Survey of Architectures, Capabilities, and Deployment Trade offs
arxiv.org·3h
⚙️TLA+
Deterministic AI: Why Reliability, Not Creativity, Is the Future of LLMs
davletd.medium.com·11h·
Discuss: Hacker News
⚙️TLA+
BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMs
arxiv.org·3h
✓Automated Theorem Proving
Automated Verification of Code Logic & Security Vulnerabilities via Hyperdimensional Semantic Analysis
dev.to·1d·
Discuss: DEV
📏Code Metrics
SECA: Semantically Equivalent and Coherent Attacks for Eliciting LLM Hallucinations
arxiv.org·3h
👑Coq Tactics
An alternative to knowledge graphs for storing loosely structured content
fleetingswallow.com·1d·
Discuss: Hacker News
🕸️Knowledge Graphs
Scaling Code-Assisted Chain-of-Thoughts and Instructions for Model Reasoning
arxiv.org·3h
⚡Incremental Computation
Aria: An Agent For Retrieval and Iterative Auto-Formalization via Dependency Graph
arxiv.org·3h
⚡Proof Automation
Towards a Typology of LLM Chains-of-Thought
1a3orn.com·12h·
Discuss: Hacker News
🌳Context free grammars
DRPO: Efficient Reasoning via Decoupled Reward Policy Optimization
arxiv.org·3h
🎯Performance Proofs