Editing 2635: Superintelligent AIs
Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision | Your text | ||
Line 31: | Line 31: | ||
::"Black Hat picks up and opens the box. A little glowy ball comes out of it."[https://xkcd.com/1450/info.0.json] | ::"Black Hat picks up and opens the box. A little glowy ball comes out of it."[https://xkcd.com/1450/info.0.json] | ||
β | While LaMDA is not the first very large {{w|language model}} based on {{w| | + | While LaMDA is not the first very large {{w|language model}} based on {{w|Transformer (machine learning model)|transformer-based machine learning}} technology which has been claimed to be sentient,[https://www.youtube.com/watch?v=PqbB07n_uQ4] it does have a variety of new characteristics beyond what those of its predecessors, such as {{w|GPT-3}} (including [https://beta.openai.com/playground/ OpenAI's Davinci]) and NVIDIA GPT-2 offshoots, include. In particular, LaMDA's {{w|deep learning}} {{w|connectionist}} {{w|neural net}} has access to multiple {{w|Symbolic systems|symbolist}} text processing systems, [https://towardsdatascience.com/why-gpt-wont-tell-you-the-truth-301b48434c2c including a database] (which apparently includes a real-time clock and calendar), a mathematical calculator, and a natural language translation system, giving it superior accuracy in tasks supported by those systems, and making it among the first {{w|Dual process theory|dual process}} chatbots. LaMDA also is not {{w|Stateless protocol|stateless}}, because its "{{w|sensibility|sensibleness}}" metric (including whether responses contradict anything said earlier) is {{w|fine-tuning|fine-tuned}} by "pre-conditioning" each dialog turn by prepending 14-30{{cn}} of the most recent dialog interactions, on a user-by-user basis.[https://arxiv.org/pdf/2201.08239.pdf [p. 6 here]] LaMDA is tuned on nine unique performance metrics, almost all of which its predecessors were not: Sensibleness, Specificity, Interestingness, Safety, Groundedness, Informativeness, Citation accuracy, Helpfulness, and Role consistency.[''ibid.,'' pp. 5-6.] |
==Transcript== | ==Transcript== |