Teresa Berndtsson / Higher Photos of AI / Letter Phrase Textual content Taxonomy / Licenced by CC-BY 4.0.
We’ve collected a few of the articles, opinion items, movies and assets regarding massive language fashions. A few of these hyperlinks additionally cowl different generative fashions. We’ll periodically replace this listing so as to add any additional assets of curiosity.
How they work
- What are Generative AI fashions?, Kate Soule, video from IBM Expertise.
- What’s GPT-4 and the way does it differ from ChatGPT?, Alex Hern, The Guardian.
- What Is ChatGPT Doing … and Why Does It Work?, Stephen Wolfram.
- Understanding Giant Language Fashions — A Transformative Studying Listing, Sebastian Raschka.
- How ChatGPT is Skilled, video by Ari Seff.
- ChatGPT – what’s it? How does it work? Ought to we be excited? Or scared?Deep Dhillon, The Radical AI podcast.
Journal, convention and arXiv articles
- Scientists’ Views on the Potential for Generative AI of their Fields, Meredith Ringel Morris, arXiv.
- LaMDA: Language Fashions for Dialog Purposes, Romal Thoppilan et al, arXiv.
- What Language Mannequin to Practice if You Have One Million GPU Hours?, Teven Le Scao et al, arXiv.
- Alpaca: A Sturdy, Replicable Instruction-Following Mannequin, Rohan Taori et al.
- Course of for Adapting Language Fashions to Society (PALMS) with Values-Focused Datasets, Irene Solaiman, Christy Dennison, NeurIPS 2021.
- On the Risks of Stochastic Parrots: Can Language Fashions Be Too Large?
, Emily Bender, Timnit Gebru, Angelina McMillan-Main, Shmargaret Shmitchell, FAccT 2021. - A Survey of Giant Language Fashions, Wayne Xin Zhao et al, arXiv.
- A Watermark for Giant Language Fashions, John Kirchenbauer, Jonas Geiping, Yuxin Wen, Jonathan Katz, Ian Miers, Tom Goldstein, arXiv.
- Between Subjectivity and Imposition: Energy Dynamics in Knowledge Annotation for Laptop Imaginative and prescient, Milagros Miceli, Martin Schuessler, Tianling Yang, Proceedings of the ACM on Human-Laptop Interplay.
- AI classifier for indicating AI-written textual content, OpenAI.
- Pythia: A Suite for Analyzing Giant Language Fashions Throughout Coaching and Scaling, Stella Biderman et al, arXiv.
- GPT-4 Technical Report, OpenAI, arXiv.
- GPT-4 System Card, OpenAI.
- BloombergGPT: A Giant Language Mannequin for Finance, Shijie Wu et al, arXiv.
Newspaper, journal, College web site, and blogpost articles
- Why exams meant for people may not be good benchmarks for LLMs like GPT-4, Ben Dickson, Enterprise Beat.
- Does GPT-4 Actually Perceive What We’re Saying?, David Krakauer, Nautilus.
- Giant language fashions are biased. Can logic assist save them?, Rachel Gordon, MIT Information.
- Ecosystems graph for ML fashions and their relationships, researchers at Stanford College.
- ChatGPT struggles with Wordle puzzles, which says so much about the way it works, Michael G. Madden, The Dialog.
- AIhub espresso nook: Giant language fashions for scientific writing, AIhub.
- ChatGPT Is a Blurry JPEG of the Net, Ted Chiang, The New Yorker.
- ChatGPT, Galactica, and the Progress Entice, Abeba Birhane and Deborah Raji, Wired.
- ChatGPT can’t deceive you, however you continue to shouldn’t belief it, Mackenzie Graham, The Dialog.
- AI info retrieval: A search engine researcher explains the promise and peril of letting ChatGPT and its cousins search the net for you, Chirag Shah, The Dialog.
- A small step for analysis however a large leap for utility, Interview with Fredrik Heintz, Linköping College.
- ChatGPT threatens language variety. Extra must be accomplished to guard our variations within the age of AI, Collin Bjork, The Dialog.
- Column: Afraid of AI? The startups promoting it need you to be, Brian Service provider, Los Angeles Instances.
- 3 ways AI chatbots are a safety catastrophe, Melissa Heikkilä, MIT Tech Evaluate.
- Time: OpenAI Used Kenyan Staff on Much less Than $2 Per Hour to Make ChatGPT Much less Poisonous, Billy Perrigo, TIME.
- Misplaced fears of an ‘evil’ ChatGPT obscure the true hurt being accomplished, John Naughton, The Guardian.
- Darktrace warns of rise in AI-enhanced scams since ChatGPT launch, Mark Sweney, The Guardian.
- Lawmakers battle to distinguish AI and human emails, Kate Blackwood, Cornell Chronicle.
- Colombian decide says he used ChatGPT in ruling, Luke Taylor, The Guardian.
- Bhashini: At your service an Indian language chatbot powered by ChatGPT, video from The Financial Instances.
Podcasts and video discussions
- The Limitations of ChatGPT with Emily M. Bender and Casey Fiesler, Radical AI Podcast.
- CLAIRE AQuA: “ChatGPT and Giant Language Fashions”, CLAIRE.
- Su Lin Blodgett on Creating Simply Language Applied sciences, The Good Robotic Podcast.
Give attention to LLMs and robotics
- ChatGPT for Robotics: Design Ideas and Mannequin Talents, Microsoft.
- OpenAI and Determine be part of the race to humanoid robotic employees, Loz Blain, New Atlas.
- Interior Monologue: Embodied Reasoning by means of Planning with Language Fashions, Wenlong Huang et al., arXiv.
- PaLM-E: An embodied multimodal language mannequin, Danny Driess, Google.
- Consciousness, Embodiment, Language Fashions (with Professor Murray Shanahan), YouTube video from Machine Studying Avenue Speak.
Give attention to LLMs and training
- Opinion: ChatGPT – what does it imply for tutorial integrity?, Giselle Byrnes, Massey College.
- Debate: ChatGPT gives unseen alternatives to sharpen college students’ important expertise, Erika Darics, Lotte van Poppel, The Dialog.
- ChatGPT and dishonest: 5 methods to vary how college students are graded, Louis Volante, Christopher DeLuca Don A. Klinger, The Dialog.
- ChatGPT: college students might use AI to cheat, however it’s an opportunity to rethink evaluation altogether, Sam Illingworth, The Dialog.
- A Trainer’s Immediate Information to ChatGPT, @herfteducator.
Regarding artwork and different artistic processes
- ‘ChatGPT stated I didn’t exist’: how artists and writers are combating again in opposition to AI, Vanessa Thorpe, The Guardian.
- AI and the way forward for work: 5 consultants on what ChatGPT, DALL-E and different AI instruments imply for artists and data employees, Lynne Parker, Casey Greene, Daniel Acuña, Kentaro Toyama Mark Finlayson, The Dialog.
- Is there a strategy to pay content material creators whose work is used to coach AI? Sure, however it’s not foolproof, Brendan Paul Murphy, The Dialog.
- ChatGPT is the push greater training must rethink evaluation, Sioux McKenna, Dan Dixon, Daniel Oppenheimer, Margaret Blackie, Sam Illingworth, The Dialog.
- AI Artwork: How artists are utilizing and confronting machine studying, YouTube video from the Museum of Fashionable Artwork.
Misinformation, pretend information and the influence on journalism
- Misinformation Monitor: March 2023, deal with GPT-4, NewsGuard.
- A pretend information frenzy: why ChatGPT may very well be disastrous for reality in journalism, Emily Bell, The Guardian.
- Defending Towards Neural Faux Information, Rowan Zellers et al, arXiv.
Regulation and coverage
- ‘Political propaganda’: China clamps down on entry to ChatGPT, Helen Davidson, The Guardian.
- Chatbots, deepfakes, and voice clones: AI deception on the market, USA Federal Commerce Fee weblog submit.
