Cryptocurrency news

The network noticed gpt2-chatbot — the probable successor of GPT-4 from OpenAI

Forklog / 30.04.2024 / 11:58
The network noticed gpt2-chatbot — the probable successor of GPT-4 from OpenAI
The gpt2 chatbot language model, which is associated with OpenAI, has appeared on the LMSYS platform.The chatbot provides more specific answers compared to the GPT-4 Turbo, and is also able to search for logical connections.Sam Altman, the head of OpenAI, drew attention to the gpt2 chatbot, although the origin of the model is still unknown.

A chatbot named gpt2-chatbot has appeared on the LMSYS Chatbot Arena Leaderboard platform for evaluating language models. It was probably developed by OpenAI.

Users who managed to interact with the model noted a significant leap forward compared to the GPT-4 Turbo. According to them, the chatbot's answers look more specific, and the tool itself is able to search for non-obvious logical and associative connections.

I found one task that gpt2-chatbot is better than all other models, and it's completely useless.
Early but rapid ascent on the A+B-1 question by @Kangwook_Lee pic.twitter.com/xwOfnB1r03

— Dimitris Papailiopoulos (@DimitrisPapail) April 29, 2024

One of the users asked gpt2-chatbot to solve the problem of the International Mathematical Olympiad. The chatbot solved it the first time, which is not always possible for already released language models.

uh.... gpt2-chatbot just solved an International Math Olympiad (IMO) problem in one-shot

the IMO is insanely hard. only the FOUR best math students in the USA get to compete

prompt + its thoughts 🧵 https://t.co/CuO0ToJmb9 pic.twitter.com/3xxWPvtmuG

— Andrew Gao (@itsandrewgao) April 29, 2024

Other users tried using a chatbot to generate Python code for the Snake game. 

This was the game it gave me! Code it right in the first try pic.twitter.com/ihMhBR9BAo

— Alvaro Cintas (@dr_cintas) April 29, 2024

Why is gpt2-chatbot associated with OpenAI

After the model appeared on LMSYS, X users began to build theories about the origin of gpt2-chatbot. The first thing that indicates a connection with OpenAI is the use of the abbreviation GPT in the name of the chatbot. It was a California startup that invented and popularized it, although it could not register it as a trademark.

One of the users claims that the tool explicitly named OpenAI as the developer of the model.

No, OpenAI pic.twitter.com/PioFY7K1a6

— Ted Punt (@punt_ted) April 29, 2024

Moreover, on the night of April 30, the head of OpenAI, Sam Altman, drew attention to gpt2-chatbot.

"I'm not indifferent to gpt2," he wrote in X.

Altman rarely comments on the rumors surrounding his company's developments. Therefore, many users considered his tweet to be a hint of the startup's involvement in gpt2-chatbot.

The successor to the GPT-4, a completely new model or an old development

Enthusiasts believe that the chatbot can be GPT-4.5 or the next version of GPT-5. Some have suggested that this is the original GPT-4 without the limits that the developers set before the release.

The number in the name can also be misleading. The original GPT-2 model was released in November 2019 and did not have the capabilities of its successors. However, there is a theory that GPT2 is the second iteration of all OpenAI models, and not the classic serial number.

Anyone can try gpt2-chatbot

While the public is waiting for news from OpenAI, anyone can independently experience the capabilities of gpt2-chatbot. To do this, go to the LMSYS website and select the Direct Chat section.

Choosing a specific chatbot on the LMSYS platform. Data: LMSYS.

After that, you need to find gpt2-chatbot in the drop-down list and go down the page to the text input field.

Search for gpt2-chatbot in the drop-down list on the LMSYS platform. Data: LMSYS.

Unfortunately, live chat has a limit of 1000 messages per hour for all users of the platform. Due to the high hype, it becomes more difficult to use it.

Limitations of gpt2-chatbot on the LMSYS platform. Data: LMSYS.

However, it can be found in the Arena (battle) mode, where two models are randomly selected for a "blind" comparison. In this case, it will take some time for the gpt2 chatbot to fall out.

Recall that in November 2023, OpenAI introduced its most current GPT-4 Turbo model.

Source
Recently News

© Token Radar 2024. All Rights Reserved.
IMPORTANT DISCLAIMER: All content provided herein our website, hyperlinked sites, associated applications, forums, blogs, social media accounts and other platforms (“Site”) is for your general information only, procured from third party sources. We make no warranties of any kind in relation to our content, including but not limited to accuracy and updatedness. No part of the content that we provide constitutes financial advice, legal advice or any other form of advice meant for your specific reliance for any purpose. Any use or reliance on our content is solely at your own risk and discretion. You should conduct your own research, review, analyse and verify our content before relying on them. Trading is a highly risky activity that can lead to major losses, please therefore consult your financial advisor before making any decision. No content on our Site is meant to be a solicitation or offer.