In this report, we introduce our first-generation reasoning model, LexPro-1.0, a large language model designed for the highly specialized Chinese legal domain, offering comprehensive capabilities to meet diverse realistic needs. Existing legal LLMs face two primary challenges. Firstly, their design and evaluation are predominantly driven by computer science perspectives, leading to insufficient incorporation of legal expertise and logic, which is crucial for high-precision legal applications, such as handling complex prosecutorial tasks. Secondly, these models often underperform due to a lack of comprehensive training data from the legal domain, limiting their ability to effectively address real-world legal scenarios. To address this, we first compile millions of legal documents covering over 20 types of crimes from 31 provinces in China for model training. From the extensive dataset, we further select high-quality for supervised fine-tuning, ensuring enhanced relevance and precision. The model further undergoes large-scale reinforcement learning without additional supervision, emphasizing the enhancement of its reasoning capabilities and explainability. To validate its effectiveness in complex legal applications, we also conduct human evaluations with legal experts. We develop fine-tuned models based on DeepSeek-R1-Distilled versions, available in three dense configurations: 14B, 32B, and 70B.

In the filing, Cohen wrote that he had not kept up with "emerging trends (and related risks) in legal technology and did not realize that Google Bard was a generative text service that, like ChatGPT, could show citations and descriptions that looked real but actually were not." To him, he said, Google Bard seemed to be a "supercharged search engine."

The backup Uber driver for a self-driving vehicle that killed a pedestrian in suburban Phoenix in 2018 pleaded guilty Friday to endangerment in the first deadly crash involving a fully autonomous car. Arizona state judge David Garbarino, who accepted the plea agreement, sentenced Rafaela Vasquez to three years of supervised probation for the crash that killed 49-year-old Elaine Herzberg. Vasquez, 49, told police that Herzberg "came out of nowhere" and that she didn't see Herzberg before hitting her on a darkened Tempe street on 18 March 2018. Vasquez had been charged with felony negligent homicide. The charge to which she pleaded could be reclassified as a misdemeanor if she completes probation. Authorities say Vasquez was streaming the television show The Voice on a phone and looking down in the moments before Uber's Volvo XC-90 SUV struck Herzberg, who was crossing with her bicycle.

Rep. Alexandria Ocasio-Cortez, speaking at an event in January 2019 honoring the legacy of Dr. Luther King, said, "Algorithms are still made by human beings, and those algorithms are still pegged to basic human assumptions. And if you don't fix the bias, then you are just automating the bias." Though her comments were correct--algorithms can indeed reflect and exhibit human bias--Rep. Ocasio-Cortez's framing of the intersection of algorithms and fairness highlighted an often-ignored issue in progressive politics. The political movement, defined in part by its commitment to social justice, is unsurprisingly critical of the potential for algorithms, particularly AI, to facilitate discrimination, yet seemingly pays little attention to the ways in which algorithms can actually reduce discrimination.

LONDON (Reuters) - A life-like avatar called Cora is being put through her paces by Royal Bank of Scotland (RBS.L), helping customers with basic queries and giving its digital banking drive a more human face. RBS, which recently axed a quarter of its branches and has announced thousands of job cuts in the past year alone as it reduces costs, will only deploy Cora to NatWest customers if the bank's female avatar passes her probation period. The digital teller, who wears a NatWest branded uniform and has an ear piercing and sparkling teeth, answers simple questions on getting a mortgage or what to do if a customer loses their card. "It could create another way for our customers to bank with us on top of the usual services we offer and be used to help answer questions round the clock, whilst cutting queuing times," Kevin Hanley, director of innovation at NatWest, said. Cora, which is undergoing advanced testing as part of a pilot program at RBS, could even be used to train members of staff, Hanley said on Wednesday.

It is fitting that the greatest code-breaker of World War Two remains a riddle a hundred years after his birth. Alan Turing, the brilliant, maverick mathematician, widely considered to be the father of computer science and artificial intelligence, invented an electromagnetic machine called the'bombe' which formed the basis for deciphering Germany's Enigma codes. The man himself has rather eluded definition: painted (too easily) as a nutty professor with a squeaky voice; as a quirky, haphazard character with a sloppy appearance by his mother and schoolmasters; by colleagues as a gruff, socially awkward man; and by his friends as an open-hearted, generous and gentle soul. The crucial contribution Turing made at Bletchley Park, one that has been credited with shortening the war by two years and saving countless lives, did not become public knowledge until twenty years after his death. His mother, brother and friends did not know until long after they'd mourned him, the extent of his heroism.