💥 Gate Square Event: #PTB Creative Contest# 💥
Post original content related to PTB, CandyDrop #77, or Launchpool on Gate Square for a chance to share 5,000 PTB rewards!
CandyDrop x PTB 👉 https://www.gate.com/zh/announcements/article/46922
PTB Launchpool is live 👉 https://www.gate.com/zh/announcements/article/46934
📅 Event Period: Sep 10, 2025 04:00 UTC – Sep 14, 2025 16:00 UTC
📌 How to Participate:
Post original content related to PTB, CandyDrop, or Launchpool
Minimum 80 words
Add hashtag: #PTB Creative Contest#
Include CandyDrop or Launchpool participation screenshot
🏆 Rewards:
🥇 1st
OpenAI Reveals the Source of LLM Hallucinations and Proposes a New Training Method to Reduce Confident Errors
【Coin World】OpenAI discovered that the hallucination phenomenon in large language models stems from their training and evaluation methods, which encourage guessing rather than admitting uncertainty. The company suggests increasing the penalties for confidential error and awarding partial points for uncertain answers, similar to negative scoring in standardized tests. Data shows that models rewarded for accuracy have higher error rates, while those that acknowledge uncertainty perform more reliably. OpenAI is implementing these improvements to reduce the hallucination phenomenon in its latest models.