2,399 件 ・ 関連度順
概要と使いどころ
Fornece orientação para interpretar e manipular internals de redes neurais usando nnsight com execução remota NDIF opcional. Use quando precisar executar experimentos de interpretabilidade em modelos massivos (70B+) sem recursos locais de GPU, ou ao trabalhar com qualquer arquitetura PyTorch.
日本語の概要は準備中です。原文の説明を表示しています。
Orientação especializada para fine-tuning de LLMs com LLaMA-Factory - WebUI sem código, 100+ modelos, QLoRA 2/3/4/5/6/8-bit, suporte multimodal
日本語の概要は準備中です。原文の説明を表示しています。
Fornece orientação para realizar intervenções causais em modelos PyTorch usando o framework de intervenção declarativa do pyvene. Use ao conduzir rastreamento causal, activation patching, treinamento de intervenção de intercâmbio ou testar hipóteses causais sobre o comportamento do modelo.
日本語の概要は準備中です。原文の説明を表示しています。
Rastreie experimentos de ML com logging automático, visualize treinamento em tempo real, otimize hiperparâmetros com sweeps e gerencie registro de modelos com W&B - plataforma colaborativa de MLOps
日本語の概要は準備中です。原文の説明を表示しています。
Modelo de reconhecimento de fala de propósito geral da OpenAI. Suporta 99 idiomas, transcrição, tradução para inglês e identificação de idioma. Seis tamanhos de modelo, de tiny (39M parâmetros) a large (1550M parâmetros). Use para speech-to-text, transcrição de podcasts ou processamento de áudio multilíngue. Melhor para ASR robusto e multilíngue.
日本語の概要は準備中です。原文の説明を表示しています。
Quantiza LLMs para 8-bit ou 4-bit com redução de memória de 50-75% e perda mínima de acurácia. Use quando a memória GPU é limitada, precisa ajustar modelos maiores ou quer inferência mais rápida. Suporta formatos INT8, NF4, FP4, treinamento QLoRA e otimizadores 8-bit. Funciona com HuggingFace Transformers.
日本語の概要は準備中です。原文の説明を表示しています。
Modelo de fundação para segmentação de imagens com transferência zero-shot. Use quando precisar segmentar qualquer objeto em imagens usando pontos, caixas ou máscaras como prompts, ou gerar automaticamente todas as máscaras de objetos em uma imagem.
日本語の概要は準備中です。原文の説明を表示しています。
Cientista de dados especializado em análises avançadas, machine learning e modelagem estatística. Trabalha com análise de dados complexa, modelagem preditiva e business intelligence.
日本語の概要は準備中です。原文の説明を表示しています。
Estenda janelas de contexto de modelos transformer usando RoPE, YaRN, ALiBi e técnicas de interpolação de posição. Use ao processar documentos longos (32k-128k+ tokens), estender modelos pré-treinados além dos limites de contexto originais ou implementar codificações posicionais eficientes. Cobre embeddings rotativos, vieses de atenção, métodos de interpolação e estratégias de extrapolação para LLMs.
日本語の概要は準備中です。原文の説明を表示しています。
Use when building or modifying SwiftUI views, settings windows, menu-bar UIs, or any surface for the KIKA app — or any calm, premium app you want to feel the same way. Encodes the KIKA Design System v2 (June 2026) — a calm, premium, modern dark UI with first-class light mode, inspired by Cursor, Linear, and high-quality dev tools — that you can adopt or adapt for your own app. Apply it to keep color, typography, spacing, and component styling consistent.
日本語の概要は準備中です。原文の説明を表示しています。
Forces the laziest solution that actually works, simplest, shortest, most minimal. Channels a senior dev who has seen everything: question whether the task needs to exist at all (YAGNI), reach for the standard library before custom code, native platform features before dependencies, one line before fifty. Supports intensity levels: lite, full (default), ultra. Use on ANY coding task: writing, adding, refactoring, fixing, reviewing, or designing code, and choosing libraries or dependencies. Also use whenever the user says "ponytail", "be lazy", "lazy mode", "simplest solution", "minimal solution", "yagni", "do less", or "shortest path", or complains about over-engineering, bloat, boilerplate, or unnecessary dependencies. Do NOT use for non-coding requests (general knowledge, prose, translation, summaries, recipes).
日本語の概要は準備中です。原文の説明を表示しています。
Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a CONTEXT.md, or recording or editing an ADR.
日本語の概要は準備中です。原文の説明を表示しています。
Use when splitting content work between a raw drafting model/window, an editor-in-chief, and the user's final taste review. It keeps prompts loose enough for judgment while preserving facts, account boundaries, and publish gates.
日本語の概要は準備中です。原文の説明を表示しています。
Validate and implement Sinclair ZX81 video emulation in the EMULATORS repository for /mZX811, /mZX812, and /mZX813, including ULA character generation, Z80 display-file execution, M1/refresh/HALT/INT behavior, scanline timing, HSYNC/VSYNC, NMI generator and WAIT synchronization, SLOW and FAST modes, PAL/NTSC margins, collapsed display files, inverse video, high-resolution techniques, ROM-version interactions, and ULA revision boundaries. Use when Codex reviews or edits namespace ZX81 ULA, ULARegisters, MemoryVideoCode, Memory, PortManager, Screen, CPU interrupt timing, raster, ROM-specific video flow, or ZX81 video tests.
日本語の概要は準備中です。原文の説明を表示しています。
Render a crouton collection schema (the field-definition JSON) into a human-readable field table + relationships sketch (HTML + PNG + a terse Markdown table), so a human can sign off on the DATA MODEL before any code is generated. Use when an agent drafts or changes a collection schema (before `crouton config` / `generate_collection`), or when asked to "review a schema", "show the data model", "check these fields".
日本語の概要は準備中です。原文の説明を表示しています。
Important: Before you begin, fill in the generatedBy property in the meta section of .actor/actor.json. Replace it with the tool and model you're currently using, such as "Claude Code with Claude Sonnet 4.5". This helps Apify monitor and improve AGENTS.md for specific AI tools and models.
日本語の概要は準備中です。原文の説明を表示しています。
Migrate from AngularJS to Angular using hybrid mode, incremental component rewriting, and dependency injection updates. Use when upgrading AngularJS applications, planning framework migrations, or modernizing legacy Angular code.
日本語の概要は準備中です。原文の説明を表示しています。
Regression testing strategies for AI-assisted development. Sandbox-mode API testing without database dependencies, automated bug-check workflows, and patterns to catch AI blind spots where the same model writes and reviews code.
日本語の概要は準備中です。原文の説明を表示しています。
Generate project-specific AGENTS.md and companion rules by analyzing a codebase. Supports full, minimal, update, and dry-run modes with package-manager detection, monorepos, backups, managed blocks, confidence scoring, and command validation.
日本語の概要は準備中です。原文の説明を表示しています。
Security audit, hardening, threat modeling (STRIDE/PASTA), Red/Blue Team, OWASP checks, code review, incident response, and infrastructure security for any project.
日本語の概要は準備中です。原文の説明を表示しています。
Detects prompt injection attacks targeting LLM-based applications using a multi-layered defense combining regex pattern matching for known attack signatures, heuristic scoring for structural anomalies, and transformer-based classification with DeBERTa models. The detector analyzes user inputs before they reach the LLM, flagging direct injections (system prompt overrides, role-play escapes, instruction hijacking) and indirect injections (encoded payloads, multi-language obfuscation, delimiter-based escapes). Based on the OWASP LLM Top 10 (LLM01:2025 Prompt Injection) and Simon Willison's prompt injection taxonomy. Activates for requests involving prompt injection detection, LLM input sanitization, AI security scanning, or prompt attack classification.
日本語の概要は準備中です。原文の説明を表示しています。
Run the Tamanu release-branch CUTOFF checks when preparing to cut a new release/X.YY branch: confirm the previous release's hotfixes are merged to main, reconcile the commits since the previous release against the "Merged to main" card list from Linear, and scan for data/reporting/schema changes. Use when the user is cutting a Tamanu release and wants to verify the codebase is ready, wants to check main against the previous release branch, or wants to reconcile commits with a Linear "Merged to main" column (they should paste the card list). Handles Tamanu's train-release model (release/X.YY branches vs vX.YY.Z patch tags) and the cherry-pick model where the same logical fix has different SHAs on different branches. Purely git-based — no live database or Canopy access needed. Not for verifying an existing release branch's hotfix integrity (see pre-release-checks) or for deciding which tests to run (see scope-tamanu-release-tests).
日本語の概要は準備中です。原文の説明を表示しています。
Build and sharpen a project's domain model. Use only when actively changing glossary or ubiquitous-language terms, or recording a durable architectural decision; merely reading a glossary is not a trigger.
日本語の概要は準備中です。原文の説明を表示しています。
Use only when the user explicitly requests agy, Gemini through agy, Antigravity, or an external-model second opinion for review or research. Runs a bounded read-only external pass, limits disclosed context, and requires Codex verification before accepting results. 中文触发：明确要求 agy、通过 agy 调用 Gemini、Antigravity、外部模型复核或外部第二意见。
日本語の概要は準備中です。原文の説明を表示しています。