Züri Bar Yochay zuriby

What is your setup?

Just chat.deepseek.com with prompts adapted from this gist.

Does it work in one-shot or I have to prompt it multiple times?

For the qX_0 variants, they are actually quite straight-forward so deepseek can come up with a correct result in 1 shot.
For the qX_K it's more complicated, I would say most of the time I need to re-prompt it 4 to 8 more times.
The most difficult was q6_K, the code never works until I ask it to only optimize one specific part, while leaving the rest intact (so it does not mess up everything)

A brief example of the latest idea for communicating with PhantomJS from the client-side.

This is the next revision of ideas discussed in a previous Gist:
https://gist.github.com/JamesMGreene/3716654

This example demonstrates shimming in the new communication channels under discussion via existing methods (WebPage#onCallback and window.callPhantom).

The example does, however, take advantage of a yet-to-be-implemented EventEmitter API for core modules in the PhantomJS outer context. This API would add the following methods to core modules as appropriate:

on / addEventListener

Basecamp was under network attack

The attack detailed below has stopped (for the time being) and almost all network access for almost all customers have been restored. We're keeping this post and the timeline intact for posterity. Unless the attack resumes, we'll post a complete postmortem within 48 hours (so before Wednesday, March 26 at 11:00am central time).

Criminals have laid siege to our networks using what's called a distributed denial-of-service attack (DDoS) starting at 8:46 central time, March 24 2014. The goal is to make Basecamp, and the rest of our services, unavailable by flooding the network with bogus requests, so nothing legitimate can come through. This attack was launched together with a blackmail attempt that sought to have us pay to avoid this assault.

Note that this attack targets the network link between our servers and the internet. All the data is safe and sound, but nobody is able to get to it as long as the attack is being successfully executed. This is like a bunch of people

	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

	# train_grpo.py
	import re
	import torch
	from datasets import load_dataset, Dataset
	from transformers import AutoTokenizer, AutoModelForCausalLM
	from peft import LoraConfig
	from trl import GRPOConfig, GRPOTrainer

	# Load and prep dataset

	#!/usr/bin/perl
	#
	# aggregate-cidr-addresses - combine a list of CIDR address blocks
	# Copyright (C) 2001,2007 Mark Suter <[email protected]>
	#
	# This program is free software: you can redistribute it and/or modify
	# it under the terms of the GNU General Public License as published by
	# the Free Software Foundation, either version 3 of the License, or
	# (at your option) any later version.
	#

	#include <stdio.h>
	#include <time.h>

	/* TODO: resume.h */

	typedef struct {
	char * company;
	char * location;
	char * title;

	M[16],X=16,W,k;main(){T(system("stty cbreak")
	);puts(W&1?"WIN":"LOSE");}K[]={2,3,1};s(f,d,i
	,j,l,P){for(i=4;i--;)for(j=k=l=0;k<4;)j<4?P=M
	[w(d,i,j++)],W\|=P>>11,l*P&&(f?M[w(d,i,k)]=l<<
	(l==P):0,k++),l=l?P?l-P?P:0:l:P:(f?M[w(d,i,k)
	]=l:0,++k,W\|=2*!l,l=0);}w(d,i,j){return d?w(d
	-1,j,3-i):4*i+j;}T(i){for(i=X+rand()%X;M[i%X]
	*i;i--);i?M[i%X]=2<<rand()%2:0;for(W=i=0;i<4;
	)s(0,i++);for(i=X,puts("\e[2J\e[H");i--;i%4\|\|
	puts(""))printf(M[i]?"%4d\|":" \|",M[i]);W-2

	www:
	requirements:
	- faye
	- jade