HN Explorer

codelion

↗ HN profile [ 1421 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. Frame: Grounding LLM Vulnerability Detection with a Sound Separation-Logic Core (lambdasec.github.io)
    1 points 0
  2. Mlx-optiq: per-layer mixed-precision LLM quantization for Apple Silicon (mlx-optiq.com)
    2 points 1
  3. A 1B humanizer that matches human writing on an AI detector (mlx-optiq.com)
    3 points 0
  4. Zero-Dependency Programming (conjure.pages.dev)
    2 points 0
  5. Scaling Pedagogical Pre-Training: From Optimal Mixing to 10B Tokens (huggingface.co)
    2 points 0
  6. From HashHop to Memory-Augmented Language Models (huggingface.co)
    2 points 0
  7. The Optimal Architecture for Small Language Models (huggingface.co)
    1 points 0
  8. Enhancing LLMs with LoRA – Standardized Recipes for Capability Enhancement (huggingface.co)
    3 points 0
  9. OpenEvolve: Teaching LLMs to Discover Algorithms Through Evolution (algorithmicsuperintelligence.ai)
    53 points 4
  10. Ellora: Enhancing LLMs with LoRA Standardized Recipes for Capability Enhancement (huggingface.co)
    1 points 0
  11. The 1B Token Challenge: Finding the Perfect Pre-Training Mix (huggingface.co)
    6 points 1
  12. The 1B Token Challenge: Finding the Perfect Pre-Training Mix (huggingface.co)
    3 points 0
  13. Pivotal Token Search (PTS): Targeting Critical Decision Points in LLM Training (huggingface.co)
    2 points 1
  14. Internal Coherence Maximization(ICM): Label-Free Unsupervised Training Framework (github.com)
    1 points 0
  15. Unsupervised Model Improvement via Internal Coherence Maximization (huggingface.co)
    1 points 0
  16. Show HN: PTS Library – Analyze LLM reasoning through "thought anchors"

    I built PTS (Pivotal Token Search), an open-source library for mechanistic interpretability analysis of language models. The core feature is generating "thought anchors" - identifying which specific sentences in a…

    2 points 0
  17. Automated Discovery of High-Performance GPU Kernels with OpenEvolve (huggingface.co)
    4 points 0
  18. Adaptive Classifier: Dynamic Text Classification with Continuous Learning (huggingface.co)
    1 points 0
  19. Show HN: DeepThink Plugin – Bring Gemini 2.5's parallel reasoning to open models

    I built an open-source plugin in OptiLLM that implements Google's "Deep Think" reasoning approach for local models like DeepSeek R1 and Qwen3. Google's recent Gemini 2.5 report introduced Deep Think - a technique where…

    4 points 0
  20. Eliciting Fine-Tuned Transformer Capabilities via Inference-Time Techniques (arxiv.org)
    1 points 0
  21. Show HN: System Prompt Learning – LLMs Learn Problem-Solving from Experience

    I built a system that lets LLMs automatically learn and improve problem-solving strategies over time, inspired by Andrej Karpathy's idea of a "third paradigm" for LLM learning. The basic idea: instead of using static…

    48 points 13
  22. Show HN: AutoThink – Boosts local LLM performance with adaptive reasoning

    I built AutoThink, a technique that makes local LLMs reason more efficiently by adaptively allocating computational resources based on query complexity. The core idea: instead of giving every query the same "thinking…

    397 points 73
  23. Show HN: OpenEvolve – open-source implementation of DeepMind's AlphaEvolve

    I've built an open-source implementation of Google DeepMind's AlphaEvolve system called OpenEvolve. It's an evolutionary coding agent that uses LLMs to discover and optimize algorithms through iterative evolution. Try…

    8 points 3
  24. Pivotal Token Search (PTS): Identify critical decision points in LLM generations (github.com)
    1 points 0
  25. Show HN: An adaptive classifier that detects hallucinations in LLM/RAG outputs (github.com)
    1 points 1