The ONE LINER

GPT-4, the newest member of the GPT family of big language models, was just released by OpenAI. (LLM). Released on 14th March 2023.

[{"selector":"#anim-b5705b42-e9a1-4cc2-bdce-222640cee1d3 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(-7.258370321155944%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-b1a7e901-e734-462f-9706-72090e4e77bc","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-7cd0cc37-06dd-45de-991c-766298b11160","keyframes":{"transform":["translate3d(133.72092%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-2e1ab7d5-1d55-4f8b-881e-1f9f0fd05b4a","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-12bc2feb-3519-46bc-b6af-e52c34d9085b","keyframes":{"transform":["translate3d(-230.27523%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] The One Liner

Greg Brockman, the president and co-founder of OpenAI, recently gave a live stream demonstration of the model's capabilities. They improved the current ChatGPT model utilizing reinforcement learning from user feedback, similar to ChatGPT.

[{"selector":"#anim-d8be17bf-6527-4624-babc-6fda8932ea80 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-172d50dd-141d-4c66-8210-c5e9dbdbd91d","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-17e97650-daaf-4d6e-bd66-bbfb21526bf9","keyframes":{"transform":["translate3d(108.65671%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-347eb7cb-b1b3-41f6-b4cc-4a84d0e5b817","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-efadfc25-5574-4614-bf1a-a90323279a5a","keyframes":{"transform":["translate3d(-233.02751%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] The One Liner

OpenAI made Evals, a framework for comparing LLMs, open-source to aid in evaluating their models.

[{"selector":"#anim-d0445262-0fc5-4688-b796-63fafde266b1 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-858f514f-e5e5-4a74-9080-84075cd54a75","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-e0088adc-593b-43dc-9de9-5803ddaa8e2b","keyframes":{"transform":["translate3d(133.33334%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-0563b8a2-cbe2-4050-bc09-5cebc29e47fc","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-d94e41df-db34-44e0-96e7-24e605a82569","keyframes":{"transform":["translate3d(-230.27523%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] The One Liner

04

[{"selector":"#anim-7833056b-778d-4793-8179-98f676e7c7d9 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-25c5f8a6-f923-4c5c-bc8e-69841116281e","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-5b3e6891-a3a8-4901-92b0-7d55e4cba2d6","keyframes":{"transform":["translate3d(119.40790%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-a4505d58-11a3-4c57-bbc6-d3e5e0195e5d","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-ebfb3a6d-3012-45b6-885f-a560712d59f2","keyframes":{"transform":["translate3d(-225.68809%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] Several more users brought up the need for improved data privacy protections from OpenAI for applications in the medical and legal fields. One possible solution would be a homomorphic encryption approach , where the GPT model runs on encrypted input. The One Liner

Following the guidelines that OpenAI has put in place, the platform will block and notify the appropriate authority if a user uploads any image related to child sexual abuse.

[{"selector":"#anim-5b9d8a8d-7cda-4c92-b30f-3871e86cdfc2 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-42379d71-8d62-48e7-89a8-4bd6ee9140e4","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-55309edd-158e-4663-874b-f5d32ab34bb6","keyframes":{"transform":["translate3d(132.55814%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-f7f0ee96-2244-47f9-a815-e0c10d06bbb3","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-cd42606d-4066-4a59-98a9-098c9304b2f3","keyframes":{"transform":["translate3d(-229.3578%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] The One Liner

Compared to ChatGPT, GPT-4 can process and produce nearly eight times as much text. As a result, the chatbot can handle more prolonged interactions because it will have a more significant "memory size."

[{"selector":"#anim-cf51151d-d6b5-415f-b367-0bf3ab0e1db1 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(0%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] [{"selector":"#anim-c42c167a-be26-4a89-af60-2fe7c42ef249","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-58d8abda-11df-4836-8555-3a172ed6ec95","keyframes":{"transform":["translate3d(122.44898%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-0de014ff-029b-400e-a209-ec035cf82687","keyframes":{"opacity":[0,1]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] [{"selector":"#anim-f2d81cf6-14bd-4065-96e5-0e0449edde35","keyframes":{"transform":["translate3d(-233.02751%, 0px, 0)","translate3d(0px, 0px, 0)"]},"delay":0,"duration":600,"easing":"cubic-bezier(0.2, 0.6, 0.0, 1)","fill":"both"}] The One Liner

[{"selector":"#anim-b5c8b9e7-6aa1-433c-a61f-5b1a44585687 [data-leaf-element=\"true\"]","keyframes":{"transform":["translate(2.025669458905618%, 0%) scale(1.5)","translate(0%, 0%) scale(1)"]},"delay":0,"duration":2000,"easing":"cubic-bezier(.3,0,.55,1)","fill":"forwards"}] The One Liner TO STAY UPDATED Subscribe Now Learn more