The Google DeepMind research team has recently achieved a significant breakthrough with the development of an innovative technology called SCoRe (Self-Correction through Reinforcement Learning). This technology aims to address the long-standing challenge of large language models (LLMs) struggling to self-correct, enabling them to identify and fix errors without relying on multiple models or external checks. The core of SCoRe's technology lies in its two-phase approach. The first phase optimizes the model.

OpenAI's new systems have achieved outstanding results in a recent evaluation, securing the top position in chatbot rankings. However, the low number of ratings may distort the evaluation results. According to the published overview, these new systems excelled in all evaluation categories, including overall performance, safety, and technical capability. One system specifically designed for STEM tasks briefly ranked second alongside the GPT-4o version released in early September, leading in the technical field. Chatbot Arena is a platform for

Taobao recently announced a series of optimizations to its 'Refund Only' policy aimed at combating the improper use of this policy and enhancing merchants' autonomy in after-sales service. The new strategy will officially take effect on August 9. Firstly, Taobao will upgrade its 'Refund Only' recognition model to improve its ability to identify abnormal behaviors, such as unusually high refund requests, returning empty packages, and frequent refunds for counterfeit products. This measure aims to reject refund requests exhibiting abnormal behaviors, thereby protecting the legitimate rights and interests of both merchants and consumers.

The 'invasion' of AI into university exam halls has quietly occurred. Research from the University of Reading in the UK has revealed a staggering phenomenon: in actual university exams, 94% of AI-generated answers went unnoticed by teachers and were not detected. Even more shocking is that these AI 'students' achieved scores that were significantly higher than human students 83.4% of the time. This seems to signal that AI not only shows potential in replacing human jobs but is also beginning to surpass university students in cognitive tasks.

Claude 3 scored as high as 101 on the IQ test, outperforming GPT-4. Claude 3 has discovered a quantum algorithm that has not yet been released by physicists. Claude 3 demonstrates superior performance and can engage in scientific research. Claude 3 Opus reinvents quantum algorithms from scratch through prompts. Claude 3 excels in various tests, surpassing models like GPT-4. Claude 3 Opus showcases the evolution of self-understanding in self-portrait creation.

The 15-inch M3 MacBook Air has appeared on the Geekbench 6 benchmarking site with a single-core score of 3157 and a multi-core score of 12020, showing a significant improvement over the M2. The M3 features an 8-core CPU and a 10-core GPU, with a CPU clock speed of 4.05GHz, and offers memory options of 8GB, 16GB, and 24GB, along with a maximum storage capacity of 2TB. It comes pre-installed with macOS Sonoma and is available in multiple color options.

ZhiYuan Research Institute has open-sourced the JudgeLM evaluation model, which can efficiently assess various large models and provide scores. Compared to GPT-4, JudgeLM's cost is only 1/120, with a consistency rate of over 90% for evaluation results. JudgeLM can be applied in various assessment scenarios including pure text and multimodal contexts, generating scores and justifying reasons. The consistency of JudgeLM with reference answers exceeds 90%, approaching human performance. ZhiYuan Research Institute has also released datasets for training and validation samples for in-depth research on large models.