Results for “paradym”

Agent A5 - Paradigm & Worldview Advisor - Philosophical foundations for research design. Covers ontology, epistemology, axiology, and methodology alignment. Use when: establishing philosophical foundations, justifying methodological choices, writing positionality statements Triggers: paradigm, 패러다임, ontology, epistemology, worldview, 세계관, philosophical foundations, 철학적 기초

Creates 3Blue1Brown-style animated explainer videos, algorithm visualizations, and math animations using Manim Community Edition.

Manim CE animations: 3Blue1Brown math/algo videos.

GLaMM: Pixel Grounding Large Multimodal Model

Prisma ORM expert for schema design, migrations, query optimization, relations modeling, and database operations. Use PROACTIVELY for Prisma schema issues, migration problems, query performance, relation design, or database connection issues.

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Jailbreak LLMs: Parseltongue, GODMODE, ULTRAPLINIAN.

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.

COYO-700M: Image-Text Pair Dataset

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

Call 100+ LLMs through a single OpenAI-compatible interface with LiteLLM — use completion/acompletion/embedding with any provider (Anthropic, OpenAI, Google, Groq, Ollama, etc.), run a proxy server for team rate-limiting and cost tracking, load-balance across providers.

Bypasses safety filters on API-served LLMs using jailbreak templates, input obfuscation, and multi-model racing.

Runs large language models locally with Ollama, including model management, custom Modelfiles, and API integration. Use for private, offline LLM inference.