GPT-4 completed many of the tests within the top 10% of the cohort, while the original version of ChatGPT often finished up in the bottom 10%.
GPT-4, the latest intelligent version of the chatbot ChatGPT, can pass high school exams and law school exams with a score of 90 percent and has new processing capabilities that were not possible in previous versions.
The creator of OpenAI shared statistics from the GPT-4 test results on March 14, revealing that it can also convert image, audio, and video input to text in addition to handling “softer directions” in a creative and reliable way heart “It passes the mock alcohol test with a score of around 10% of the winners,” OpenAI added. “In contrast, the GPT-3.5 score is in the bottom 10 percent.”
The statistics show GPT-4 scored 163 in the 88th percentile on the LSAT test – the test students must pass in the United States to gain admission to law school. A GPT4 score will put you in a good position to get into one of the top 20 law schools or just a few points from the score reported to be required for admission into these schools and famous as Harvard, Stanford, Princeton or Yale.
The previous version of ChatGPT scored only 149 on the LSAT, placing it in the bottom 40%. GPT-4 also scored 298 out of 400 on the Uniform Bar Exam – the test of recently graduated law students qualified to practice as an attorney in any US jurisdiction.
The old version of ChatGPT struggled in this test, finishing in the bottom 10% with a score of 213 out of 400. When it comes to SAT Evidence-based Reading & Writing and SAT Math taken by US high school students to measure their college readiness, GPT-4 scored 93 and 89th percentile, respectively.
GPT-4 also did well in the “power” sciences, with well above average percentages in AP Biology (85-100%), Chemistry (71-88%) and Physics 2 (66-84%). However, his AP Calculus score was pretty good, ranging from the 43rd to the 59th percentile.
Another area lacking in GPT-4 is the English writing test, showing scores between the 8th and 44th percentile on two separate tests. OpenAI said that GPT-4 and GPT-3.5 passed the following tests in the 2022-2023 practice tests, and that “no special training” was used with language processing tools:
“We did not follow any specific training for these tests. Minor issues during testing were observed by the model during training, but we believe the results are representative.”
The result also caused consternation in the Twitter community. Nick Almond, the founder of FactoryDAO, told his 14,300 Twitter followers on March 14 that GPT4 will “scare people” and “disrupt” the world’s education system.
Former Coinbase CEO Conor Grogan said he embedded a live Ethereum smart contract into GPT-4, and the chatbot immediately pointed out several “security vulnerabilities” and explained how to exploit the code.
An earlier smart contract review on ChatGPT revealed that its original version was also able to detect code bugs to a reasonable extent. Rowan Cheung, founder of AI magazine The Rundown, shared a video of GPT copying fake websites hand-drawn on paper as code.
Get $200 Free Bitcoins every hour! No Deposit No Credit Card required. Sign Up