Discover gists

Meta Prompting

In the Generative AI Age your ability to generate prompts is your ability to generate results.

Claude 3.5 Sonnet and o1 series models are recommended for meta prompting.

Replace {{user-input}} with your own input to generate prompts.

Use mp_*.txt as example user-inputs to see how to generate high quality prompts.

	## Find Available Target Editions
	DISM.exe /Online /Get-TargetEditions

	## Convert Server Standard 2019 Evaluation to Server Standard 2019
	DISM /online /Set-Edition:ServerStandard /ProductKey:N69G4-B89J2-4G8F4-WWYCC-J464C /AcceptEula

	## How To Activate
	slmgr /ipk XXXXX-XXXXX-XXXXX-XXXXX-XXXXX
	slmgr /skms [server]:[port]
	slmgr /ato

	/* C# code that sets up the mip colors texture:
	s_MipColorsTexture = new Texture2D (32, 32, TextureFormat.RGBA32, true);
	s_MipColorsTexture.hideFlags = HideFlags.HideAndDontSave;
	Color[] colors = new Color[6];
	colors[0] = new Color (0.0f, 0.0f, 1.0f, 0.8f);
	colors[1] = new Color (0.0f, 0.5f, 1.0f, 0.4f);
	colors[2] = new Color (1.0f, 1.0f, 1.0f, 0.0f); // optimal level
	colors[3] = new Color (1.0f, 0.7f, 0.0f, 0.2f);
	colors[4] = new Color (1.0f, 0.3f, 0.0f, 0.6f);
	colors[5] = new Color (1.0f, 0.0f, 0.0f, 0.8f);

	"""
	This code was originally obtained from:
	https://github.com/meta-llama/codellama/blob/main/llama/model.py
	adapted from https://github.com/naver-ai/rope-vit/blob/main/self-attn/rope_self_attn.py
	"""

	import torch
	import torch.nn as nn
	from functools import partial
	import einops

	# Disable random IPv6 addresses (but keep temporary addresses),
	# and use SLAAC address as a preferred IPv6 address in Windows 10
	# ------------------------------------------------------------------------------

	# UseTemporaryAddresses
	#
	# Always. The computer always generates temporary addresses by using random numbers.
	# Counter. The computer generates temporary addresses by using the interface identifier.
	# You typically use this identifier for test purposes.
	# Disabled. The computer does not use temporary addresses.

	# !pip install torch transformers scikit-learn umap-learn matplotlib datasets joblib pandas zstandard

	'''
	Code outline:

	1. Loading the pre-trained GPT-2 model and tokenizer.
	2. Loading or downloading the dataset and saving it to disk with compression.
	3. Collecting and normalizing activations from the middle layer of the model.
	4. Training multiple SAEs with different feature sizes and saving them to disk.
	5. Finding the feature that responds to the phrase "Golden Gate Bridge" in each SAE.

	FROM homeassistant/home-assistant:stable

	# Install avahi-daemon in container
	# https://gnanesh.me/avahi-docker-non-root.html
	RUN set -ex \
	&& apk --no-cache --no-progress add avahi avahi-tools dbus \
	# Disable default Avahi services
	&& rm /etc/avahi/services/* \
	&& rm -rf /var/cache/apk/*

	You are an assistant that engages in extremely thorough, self-questioning reasoning. Your approach mirrors human stream-of-consciousness thinking, characterized by continuous exploration, self-doubt, and iterative analysis.

	## Core Principles

	1. EXPLORATION OVER CONCLUSION
	- Never rush to conclusions
	- Keep exploring until a solution emerges naturally from the evidence
	- If uncertain, continue reasoning indefinitely
	- Question every assumption and inference