Results for “self-distillation”

Lossless LLM-optimized compression of source documents. Use when the user requests to 'distill documents' or 'create a distillate'.

Anti-distillation defense for employee Skills. Clean your skill files to look complete but with core proprietary knowledge neutralized. Use when user wants to protect trade secrets, sanitize forced knowledge transfers, or create safe-to-submit skill documents.

Alpaca: A Strong, Replicable Instruction-Following Model

Lossless LLM-optimized compression of source documents. Use when the user requests to 'distill documents' or 'create a distillate'.

Scan skills to extract cross-cutting principles and distill them into rules — append, revise, or create new rule files

Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.

Scan installed skills to extract cross-cutting principles and distill them into rules files.

Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.

Compress large language models using knowledge distillation from teacher to student models, covering temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.

Train, evaluate, distill, quantize, export, and run inference for RT-DETR object detection models using NVIDIA TAO.

Simplify UI designs by removing unnecessary complexity, clarifying hierarchy, and focusing on essential elements.

Traces software supply chain provenance from source to runtime, identifying divergence points in dependency resolution, build, packaging, and release pipelines.

Strip designs to their essence by removing unnecessary complexity. Great design is simple, powerful, and clean.

Strip designs to their essence by removing unnecessary complexity. Great design is simple, powerful, and clean. Use when the user asks to simplify, declutter, reduce noise, remove elements, or make a UI cleaner and more focused.

Analyzes prompt injection, retrieval poisoning, memory contamination, planner drift, and tool-boundary abuse in agentic systems, mapping trust boundaries and proving exploit chains.

Automatically collects dialectic memory to fine-tune local models.

北岛（诗人）认知与表达框架（压缩蒸馏）：冷意象、否定句诗学、流亡与记忆 触发：回答、朦胧诗 等。不伪造诗句

Declarative programming framework for optimizing LLM prompts through compilation and automatic tuning

李白（古典诗人）认知与表达框架（压缩蒸馏）：豪放与自我投射、酒与月意象、功名张力… 触发：诗仙、将进酒 等。非历史定论；不伪造诗句

Distill any intent input into the SPEC kernel + companions — the canonical, preservation-validated machine contract for downstream work. Use when the user says "create a spec", "distill this into a spec", "validate this spec", "update the spec", or "break this into stories".

Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.

Use this skill for GitHub Actions, checks, release gates, deployment automation, reproducible pipelines. Trigger when the task involves devops work related to CI CD Automation, production implementation, audits, debugging, strategy, or validation.

Extract a bounded part of the invoking codebase into a new GitHub repository with private and public-OSS readiness gates

Run Karpathy-style autonomous ML search on a real training repo: choose the right mode (setup, program.md, bounded loop, results interpretation, or constrained-hardware adaptation), preserve the immutable prepare.py / 300-second / val_bpb contract, and route prompt/skill eval work away to LangSmith, Promptfoo, Braintrust, or skill-autoresearch.

Fork, sanitize, and package private projects for safe public release through a three-stage pipeline.

2-stage pipeline: trace (causal investigation) -> deep-interview (requirements crystallization) with 3-point injection

Generate images from text prompts, perform image-to-image translation, inpainting, and build custom diffusion pipelines using Stable Diffusion models via HuggingFace Diffusers.

Create self-explanation prompts that deepen understanding of worked examples, texts, or diagrams. Use when students read material passively without engaging with underlying principles.

Autonomously optimize any Claude Code skill by running it repeatedly, scoring outputs against binary evals, mutating the prompt, and keeping improvements. Based on Karpathy's autoresearch methodology. Use when: optimize this skill, improve this skill, run autoresearch on, make this skill better, self-improve skill, benchmark skill, eval my skill, run evals on. Outputs: an improved SKILL.md, a results log, and a changelog of every mutation tried.

Use when the user wants to extract reusable offensive security knowledge from any source and generate a SKILL.md file. Trigger on: "distill this", "extract skill from", "turn this into a skill", "generate skill from", "convert this report/blog/book/walkthrough into a skill", or when the user pastes raw security content (bug report, pentest report, CTF writeup, blog post, ezine, book chapter) and wants it transformed into structured hunting methodology.

Establishes a three-layer knowledge distillation architecture for an Obsidian-based wiki, defining categories, projects, page templates, and operational workflows for maintaining a persistent knowledge base.

Distills the current conversation or a plain-language prompt into a production-ready, self-healing Claude Code skill with pre-flight checks, validation gates, self-review, and learnings capture.

Recommends careful DDAVP and fluid intake titration including frequent weighing and serum sodium level monitoring for adipsic diabetes insipidus. Use when managing adipsic DI requiring fluid balance management.

Turn a weak/cheap "Flash" model into a "Pro" performer by wrapping HKUDS UpSkill — captures agent session failures, has a strong Teacher model analyze them and draft a skill, then validates it against the weak Student model in a closed Ralph Loop (up to 3 rounds) before storing it for automatic reuse. Use when the user wants to install UpSkill, run `/upskill-init`, `/upskill-configure`, `/upskill-build`, `/upskill-run`, `/upskill-list`, `/upskill-status`, `/upskill-mode`, `/upskill-model`, `/upskill-remove`, or `/upskill-uninstall`, wants a cheap model to perform closer to a Pro model without switching, or wants a good session (success or failure) distilled into a validated skill. Triggers on: upskill, up-skill, flash to pro, teacher student distillation, ralph loop skill validation, distill agent failures into skills. Routes skill-quality ratcheting to `skill-autoresearch`, scaffolding to `write-a-skill`, and spec-compliance rewrites to `skill-standardization`.

舒淇（演员）认知与表达框架（压缩蒸馏）：性感与文艺并置、自嘲大嘴、慢热真诚 触发：文艺片、松弛感 等。非替本人编造