Artificial intelligence can do remarkable things, but how much does it actually understand about the world? Researchers have found that surprisingly specific information can be extracted from the inner workings of large language models (LLMs) like ChatGPT. Using mathematical tools to probe a model’s internal activity, they can identify where a city is located on a map, when a historical figure lived or what emotion a character feels, even detecting descriptions of pain.Artificial intelligence can do remarkable things, but how much does it actually understand about the world? Researchers have found that surprisingly specific information can be extracted from the inner workings of large language models (LLMs) like ChatGPT. Using mathematical tools to probe a model’s internal activity, they can identify where a city is located on a map, when a historical figure lived or what emotion a character feels, even detecting descriptions of pain.[#item_full_content]

It’s no secret that an ever-increasing number of scientists use AI to help with their papers. While tasks such as polishing text are generally permissible if declared, using the technology to write papers, or parts of papers, from scratch and passing them off as original work is not. But it’s happening.It’s no secret that an ever-increasing number of scientists use AI to help with their papers. While tasks such as polishing text are generally permissible if declared, using the technology to write papers, or parts of papers, from scratch and passing them off as original work is not. But it’s happening.[#item_full_content]

Inside the human brain, billions of individual neurons connect to form a vast network capable of rapidly processing information and electrical signals. A team of researchers and engineers at Penn State is taking inspiration from these natural computers to build soft networks of water-based droplets, connected by membranes like those found in the walls of living cells.Inside the human brain, billions of individual neurons connect to form a vast network capable of rapidly processing information and electrical signals. A team of researchers and engineers at Penn State is taking inspiration from these natural computers to build soft networks of water-based droplets, connected by membranes like those found in the walls of living cells.[#item_full_content]

When humans write sentences or solve math problems, they are thinking symbolically—each number, variable or word is expressed by a symbol. While AI chatbots do not “think” that way, if prompted, they can offer cogent advice about writing, math and many other topics as if a person were offering it.When humans write sentences or solve math problems, they are thinking symbolically—each number, variable or word is expressed by a symbol. While AI chatbots do not “think” that way, if prompted, they can offer cogent advice about writing, math and many other topics as if a person were offering it.[#item_full_content]

Before large language models (LLMs) can process text, they split it into chunks such as words or word fragments. The LLM behind ChatGPT, for example, splits “LMU München” into the three chunks “LM,” “U” and “München,” so it never directly sees the individual letters in “München.” This step, which is known as subword tokenization, makes LLMs efficient but causes a range of problems, such as limited character-level understanding.Before large language models (LLMs) can process text, they split it into chunks such as words or word fragments. The LLM behind ChatGPT, for example, splits “LMU München” into the three chunks “LM,” “U” and “München,” so it never directly sees the individual letters in “München.” This step, which is known as subword tokenization, makes LLMs efficient but causes a range of problems, such as limited character-level understanding.[#item_full_content]

When new information is continually added and old information is removed, an AI system may fail to find the correct material even when that material exists. KAIST researchers have developed a vector database technology that preserves the search paths leading to the needed information as data changes. This allows AI to find up-to-date information more accurately and quickly.When new information is continually added and old information is removed, an AI system may fail to find the correct material even when that material exists. KAIST researchers have developed a vector database technology that preserves the search paths leading to the needed information as data changes. This allows AI to find up-to-date information more accurately and quickly.[#item_full_content]

Ask an AI chatbot for some medical advice and it might recommend you head to the hospital when all you really have is indigestion. One of the biggest issues with AI chatbots today is that they can still share false information, according to Northeastern experts.Ask an AI chatbot for some medical advice and it might recommend you head to the hospital when all you really have is indigestion. One of the biggest issues with AI chatbots today is that they can still share false information, according to Northeastern experts.[#item_full_content]

Interactions between humans and dogs involve a continuous exchange of physical and verbal cues. A person may pet a dog, use a hand gesture, give a verbal command or call the pet from a distance, prompting changes in the dog’s posture, position or movement. Understanding these human–animal interactions requires data that captures both participants simultaneously.Interactions between humans and dogs involve a continuous exchange of physical and verbal cues. A person may pet a dog, use a hand gesture, give a verbal command or call the pet from a distance, prompting changes in the dog’s posture, position or movement. Understanding these human–animal interactions requires data that captures both participants simultaneously.[#item_full_content]

In the production of video games and animated movies, directors and animators are constantly fine-tuning how a character’s movements should appear. In many cases, directors require subtle changes, such as making a character appear a little more restrained or a bit more expressive, and communicate subjectively how much change is needed.In the production of video games and animated movies, directors and animators are constantly fine-tuning how a character’s movements should appear. In many cases, directors require subtle changes, such as making a character appear a little more restrained or a bit more expressive, and communicate subjectively how much change is needed.[#item_full_content]

A new AI system that excels at challenging games with hidden information could someday help human decision-makers select strategies to outfox opponents in complicated situations like military maneuvers.A new AI system that excels at challenging games with hidden information could someday help human decision-makers select strategies to outfox opponents in complicated situations like military maneuvers.[#item_full_content]