Denis Shiryaev 💙💛(@literallydenis) 's Twitter Profile Photo

I was skeptical about the GPT2 chatbot, but it is undoubtedly more capable than opensource models and, in some cases, better than GPT4-turbo

But it is not better than Opus in my experience – curious to know what is behind it.

Also, about the gpt2-chatbot:
It does not have a

account_circle
JorgeIHernandez(@JorgeIHernandez) 's Twitter Profile Photo

vs ... business plan writing. Overall gpt2 have more detail in the answer and a more valuable insights IMHO. The structure of the gpt2 plan is more accurate to what a complete business plan could be.

#GPT2Chatbot vs #gpt4turbo... business plan writing. Overall gpt2 have more detail in the answer and a more valuable insights IMHO. The structure of the gpt2 plan is more accurate to what a complete business plan could be.
account_circle
Sambhav Gupta(@sambhavgupta6) 's Twitter Profile Photo

could it be that OpenAI intentionally withholding releasing a gpt model significantly ahead of its competitors to keep their cards hidden?

strategy might be to release just enough update to maintain a narrow lead at the top (like with new gpt4 turbo)

account_circle
cybersapient_official(@CyberSapient1) 's Twitter Profile Photo

Making a plugin to copy using

- Uses gpt4-turbo to generate 3 suggestions
- Automatically detects text length and context
- Instantly applies without closing window
- Custom prompts for generating names, locations, languages, etc

It’s 100% free.

account_circle
Alan Kooper(@alan_kooper) 's Twitter Profile Photo

Spustil jsem lokálně LLAMA 3 8b vs. GPT. Naschvál v tento nejvíce exponovaný čas, třeba dopoledne nebo v noci i 2x rychleji. Nejdou nahrávat videa tak jen obrázek. 3b to zvládnul asi 3x rychleji než GPT4turbo ale v mnohem horší kvalitě. Ten výstup LLAMA žádná sláva a 70b by byl

Spustil jsem lokálně LLAMA 3 8b vs. GPT. Naschvál v tento nejvíce exponovaný čas, třeba dopoledne nebo v noci i 2x rychleji. Nejdou nahrávat videa tak jen obrázek. 3b to zvládnul asi 3x rychleji než GPT4turbo ale v mnohem horší kvalitě. Ten výstup LLAMA žádná sláva a 70b by byl
account_circle
石川陽太 Yota Ishikawa(@ytiskw) 's Twitter Profile Photo

まじかよ中国のSenseNova 5.0というLLMがいきなり最新のGPT4-turboをほぼ全項目でぶち抜いたらしい

まじかよ中国のSenseNova 5.0というLLMがいきなり最新のGPT4-turboをほぼ全項目でぶち抜いたらしい
account_circle
Qian Huang(@qhwang3) 's Twitter Profile Photo

Could agents driven by powerful language models perform machine learning experimentation effectively? Our MLAgentBench paper is updated on arxiv!
arxiv.org/pdf/2310.03302

Now we include more results from claude v3 Opus, gpt4 turbo, mixtral and gemini pro! Try out MLAgentbench

Could agents driven by powerful language models perform machine learning experimentation effectively? Our MLAgentBench paper is updated on arxiv!
arxiv.org/pdf/2310.03302

Now we include more results from claude v3 Opus, gpt4 turbo, mixtral and gemini pro! Try out MLAgentbench
account_circle
Sully(@SullyOmarr) 's Twitter Profile Photo

Heres the same prompt for gpt4 turbo, llama3 and gemini 1.5 pro

GPT4-Turbo gets it right sometimes. gpt2 gets it right everytime

Heres the same prompt for gpt4 turbo, llama3 and gemini 1.5 pro

GPT4-Turbo gets it right sometimes. gpt2 gets it right everytime
account_circle
ZN 🌏(@ZNConsulting) 's Twitter Profile Photo

Introducing OpenAI's GPT-4 Turbo! 🚀💫

Enhancing ChatGPT for paid users, will this update boost math, logic, coding, & writing abilities processing up to 300 pages?

It promises crisper conversations, integrates with Microsoft’s CoPilot AI & DALL-E 3. 🌐

Introducing OpenAI's GPT-4 Turbo! 🚀💫

Enhancing ChatGPT for paid users, will this update boost math, logic, coding, & writing abilities #OneStepAhead processing up to 300 pages?

It promises crisper conversations, integrates with Microsoft’s CoPilot AI & DALL-E 3. 🌐 #GPT4Turbo
account_circle
プライオン(@plionplion) 's Twitter Profile Photo

gpt2-chatbotとGPT4-Turbo-2024-04-09をChatbotArenaで捕まえられたので、ネオチバの設定作成対決させた(創造的タスクテスト)。正直甲乙つけがたい結果でした。|ailora will プライオン note.com/willplion/n/n9…

account_circle
Phil(@phill__1) 's Twitter Profile Photo

JB Opus has no clue, gpt4-turbo makes something up on how it works and gpt2-chatbot is right on. This is a function of a quite unknown addon of a small ERP system

@JasonBotterill3 Opus has no clue, gpt4-turbo makes something up on how it works and gpt2-chatbot is right on. This is a function of a quite unknown addon of a small ERP system
account_circle
Charles Floate 📈(@Charles_SEO) 's Twitter Profile Photo

Here is the end output using GPT4-Turbo, which took under 2 minutes to generate a 3,908 word article! 👀

All formatting, optimization elements etc are done by Cuppa with some additional advanced prompting - The only other thing I would have usually added is a manipulation list

Here is the end output using GPT4-Turbo, which took under 2 minutes to generate a 3,908 word article! 👀

All formatting, optimization elements etc are done by Cuppa with some additional advanced prompting - The only other thing I would have usually added is a manipulation list
account_circle
Max ᯅ e/acc(@maxxrubin_) 's Twitter Profile Photo

Ángel (e/acc) Chubby♨️ UPDATE: It’s GPT4 Turbo with vision. December 2023 cutoff (so current GPT-4V turbo we use in ChatGPT)!

I got it to read me its system prompt. This version has NONE of the normal openAI system prompt shenanigans that typically lower performance for safety. That’s my guess

account_circle
Burny — Effective Omni(@burny_tech) 's Twitter Profile Photo

As always, my first question to this maybe GPT4.5, and I'd say it's better output compared to GPT4: Turbo

Me: Create a gigantic detailed map of as much of formal and natural sciences as possible

Maybe GPT4.5:
Creating a gigantic, detailed map encompassing the vast array of both

As always, my first question to this maybe GPT4.5, and I'd say it's better output compared to GPT4: Turbo

Me: Create a gigantic detailed map of as much of formal and natural sciences as possible

Maybe GPT4.5:
Creating a gigantic, detailed map encompassing the vast array of both
account_circle
JorgeIHernandez(@JorgeIHernandez) 's Twitter Profile Photo

vd ... diet and workout. Again, gpt2 shows detailed answers and a better way to expose the solution. I can smell internal improvements more than a new model. (more experts in the mix?)

#GPT2Chatbot vd #gpt4turbo... diet and workout. Again, gpt2 shows detailed answers and a better way to expose the solution. I can smell internal improvements more than a new model. (more experts in the mix?)
account_circle
Meng To(@MengTo) 's Twitter Profile Photo

I made a Figma plugin to rewrite UI copy using AI
figma.com/community/plug…

- Uses gpt4-turbo to generate 3 suggestions
- Automatically detects text length and context
- Instantly applies without closing window
- Custom prompts for generating names, locations, languages, etc

It’s

account_circle