333K subscribers • 4 videos on Hallway
Intermediate. Assumes working knowledge.
1M views4 years ago3 recommendations
AI and machine learningTransformers and LLMsAI safety and impact
Yannic Kilcher explains how he trained a language model on 4chan posts and analyzed its behavior when deployed back onto the platform.
#gpt4chan #4chan #ai GPT-4chan was trained on over 3 years of posts from 4chan's "politically incorrect" (/pol/) board. (and no, this is not GPT-4) EXTRA VIDEO HERE: youtube.com/watch?v=dQw4w9WgXcQ Website (try the model here): gpt-4chan.com Model (no longer available): huggingface.co/ykilcher/gpt-4chan Code: github.com/yk/gpt-4chan-public Dataset: zenodo.org/record/3606810#.YpjGgexByDU OUTLINE: - Intro - Disclaimers - Elon, Twitter, and the Seychelles - How I trained a language model on 4chan posts - How good is this model? - Building a 4chan bot - Something strange is happening - How the bot got unmasked - Here we go again - Final thoughts ERRATA: - I stated that the model is better on the automated parts of TruthfulQA than any other GPT out there, which is incorrect. There exist some small GPT-models with similar performance, I was mainly talking about the flagship models, such as GPT-3 and GPT-J. Links: Merch: ykilcher.com/merch TabNine Code Completion (Referral): bit.ly/tabnine-yannick YouTube: youtube.com/c/yannickilcher Twitter: twitter.com/ykilcher Discord: ykilcher.com/discord BitChute: bitchute.com/channel/yannic-kilcher LinkedIn: linkedin.com/in/ykilcher BiliBili: space.bilibili.com/2017636191 If you want to support me, the best thing to do is to share out the content :) If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this): SubscribeStar: subscribestar.com/yannickilcher Patreon: patreon.com/yannickilcher Bitcoin (BTC): bc1q49lsw3q325tr58ygf8sudx2dqfguclvngvy2cq Ethereum (ETH): 0x7ad3513E3B8f66799f507Aa7874b1B0eBC7F85e2 Litecoin (LTC): LQW2TRyKYetVC8WjFkhpPhtpbDM4Vw7r9m Monero (XMR): 4ACL8AGrEo5hAir8A9CeVrW8pEauWvnp1WnSDZxW7tziCDLhZAGsgzhRQABDnFy8yuM9fWJDviJPHKRjV4FWt19CJZN9D4nWhat people say. Shared as a fascinating and controversial experiment demonstrating how fine-tuned language models can mimic specific online communities.
benreesmanon Hacker News • Recommends
On “Breaking the 4Chan CAPTCHA”
thegreatdukdon Hacker News • Recommends
On “GPT-4chan”
sigmoid10on Hacker News • Recommends
On “Reddit Launches Long-Awaited IPO with $748M Target”
albertzeyeron Hacker News • Mentions
On “GPT-4chan”
sitkackon Hacker News • Mentions
On “Ask HN: GPT-3 reveals my full name – can I do anything?”
benreesmanon Hacker News • Mentions
On “Are language models deprived of electric sleep?”
visargaon Hacker News • Mentions
On “CICERO: An AI agent that negotiates, persuades, and cooperates with people”
kcmastrpcon Hacker News • Mentions
On “Conservatives Are Panicking About AI Bias, Think ChatGPT Has Gone 'Woke'”
type0on Hacker News • Mentions
On “Ask HN: Does the HN commentariat have a reductive view of what a human being is?”
flangola7on Hacker News • Mentions
On “Microsoft Kosmos-1: A Multimodal Large Language Model”
EGregon Hacker News • Mentions
On “The genie escapes: Stanford copies the ChatGPT AI for less than $600”
AlbertoGPon Hacker News • Mentions
On “Mozilla.ai: Investing in Trustworthy AI”
ProllyInfamouson Hacker News • Mentions
On “ChatGPT Generating Racist Content”
IvanAchlaqullahon Hacker News • Mentions
On “Phi-3 Technical Report”
benreesmanon Hacker News • Mentions
On “DeepSeek's Hidden Bias: How We Cut It by 76% Without Performance Loss”
sunaookamion Hacker News • Mentions
On “4chan will refuse to pay daily online safety fines, lawyer tells BBC”