GPT-4 Laziness Outbreak! Completion Rate Drops by Nearly 25% in Code Comparison Tasks

站长之家
31
Translation:
Recently, the "laziness" issue of GPT-4 has once again garnered attention. Netizens have noticed that in code comparison tasks, GPT-4's completion rate has dropped by nearly a quarter. Although Altman stated that improvements are expected in the new year, users remain troubled by its performance and optimization strategies. This phenomenon may ease somewhat in the new year, but specific improvement measures have yet to be clarified.
© Copyright AIbase Base 2024, Click to View Source - https://www.aibase.com/news/5236