I replicated Ng's RYS method and found that duplicating 3 specific layers in Qwen2.5-32B boosts reasoning by 17% and duplicating layers 12-14 in Devstral-24B improves logical deduction from 0.22→0.76 on BBH — no training, no weight changes, just routing hidden states through the same circuit twice. Tools included. Two AMD GPUs, one evening.
☆244Mar 20, 2026Updated 7 months ago
Alternatives and similar repositories for llm-circuit-finder
Users that are interested in llm-circuit-finder are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Agent learns to become the worlds top expert on SAT☆183Jun 16, 2026Updated 4 months ago
- Context Gateway is an agentic proxy that enhances any AI agent workflow with instant history compaction and context optimization tools☆647Oct 21, 2026Updated 2 weeks ago
- Bayesian CausalImpact for Python — R-compatible Gibbs sampler in Rust (PyO3), 30x faster, no TensorFlow☆20Mar 26, 2026Updated 7 months ago
- Speedrunning LoRA fine-tuning: frozen task, frozen hardware, public wall-clock leaderboard. modded-nanogpt for fine-tuning.☆151Sep 12, 2026Updated last month
- Ollama for classical ML models. AOT compiler that turns XGBoost, LightGBM, scikit-learn, CatBoost & ONNX models into native C99 inference…☆690Apr 16, 2026Updated 6 months ago
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- ☆479Updated this week
- Run Qwen Flash Next on a Mac Mini m5 64GB box☆18Oct 24, 2026Updated 2 weeks ago
- Self-hosted malware static triage platform powered by LLMs☆27Aug 19, 2026Updated 2 months ago
- The official code repo of Alignment Whack-a-Mole: Finetuning Activates Verbatim Recall of Copyrighted Books in Large Language Models☆101Oct 24, 2026Updated 2 weeks ago
- SSH auth via browser passkeys (Touch ID / Face ID). No passwords, no SSH keys — just biometrics.☆28Mar 22, 2026Updated 7 months ago
- Sub-millisecond VM sandboxes for AI agents via copy-on-write forking☆2,468Oct 18, 2026Updated 2 weeks ago
- TurboQuant WASM SIMD vector compression — 3 bits/dim with fast dot product. Requires relaxed SIMD (Chrome 114+, Firefox 128+, Safari 18+,…☆323Apr 19, 2026Updated 6 months ago
- CPU-first LLM inference server on llama.cpp. Runs useful models on free-tier ARM boxes; rewriting the input made it ~6x faster and more a…☆114Oct 12, 2026Updated 3 weeks ago
- Local-first multi-runtime AI cockpit (CLI + desktop + MCP). War-rooms, review receipts, open multi-model coding benchmarks. MIT.☆34Sep 7, 2026Updated 2 months ago
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- A language model that lives in a boot sector☆99Oct 31, 2026Updated last week
- A Python framework for self-hosted LLM tool-calling and multi-step agentic workflows☆2,265Updated this week
- Running a large language model on a PlayStation 2☆52Mar 21, 2026Updated 7 months ago
- Deterministic harness-optimization simulation on five toy tasks. Scripted agent and pattern proposer; Islo backend stub.☆28Oct 22, 2026Updated 2 weeks ago
- ☆21Mar 26, 2026Updated 7 months ago
- Unified Liminal World Editor, Procedural Engine & Game Creation Toolkit inspired by Radiant and Hammer. Optimised for low-power CPUs.☆150Updated this week
- Modern Norton Commander dual-pane file manager for the web☆15Mar 25, 2026Updated 7 months ago
- The most advanced fitness web app in history and a lot of drama inside the code.☆50Mar 23, 2026Updated 7 months ago
- Py library to generate SVG visualizations of scientific data☆30Sep 9, 2026Updated last month
- 1-Click AI Models by DigitalOcean Gradient • AdDeploy popular AI models on DigitalOcean Gradient GPU virtual machines with just a single click. Zero configuration with optimized deployments.
- ☆196May 5, 2025Updated last year
- An AI-powered RSS reader that prioritizes signal over noise for engineers and researchers.☆36Oct 30, 2026Updated last week
- Tmux for claude code☆30Jun 30, 2026Updated 4 months ago
- Tokenflood is a load testing framework for simulating arbitary loads on instruction-tuned LLMs☆47Oct 21, 2026Updated 2 weeks ago
- Turn your files into a memory filesystem for AI agents.☆43Apr 15, 2026Updated 6 months ago
- layerleak the Docker Hub Secret Scanner☆44Oct 30, 2026Updated last week
- State machine guardrails for AI agents☆512Updated this week
- NUMA-Aware Contention-Free Dynamically-Auto-Tuning Bash-Native Streaming Parallelization Engine☆364Updated this week
- Determine mesh rotations to minimise overhangs for 3D printing☆70Aug 8, 2024Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Running a big model on a small laptop☆4,488Mar 19, 2026Updated 7 months ago
- Adaptive Test-time Learning and Autonomous Specialization☆2,111Updated this week
- Persistent memory plugin for Claude Code. Tiered memory system with write gates, correction propagation, and slash commands.☆202Feb 12, 2026Updated 8 months ago
- Information-dense, coverage-guided binary fuzzer☆45Updated this week
- Running PrismML's 1-bit Bonsai Image model (FLUX.2 klein 4B) on a 2025 Android flagship's Hexagon NPU. The Android counterpart to bonsai-…☆13Jun 30, 2026Updated 4 months ago
- Download arxiv PDFs from PaSa search results for local AI-powered literature analysis.☆59Apr 1, 2026Updated 7 months ago
- An open source Cloudflare Worker app and API to check if email addresses are throwaway/disposable or valid☆48Oct 29, 2026Updated last week