Carnegie Mellon University researchers propose a new LLM training technique that gives developers more control over chain-of-thought length.

Manus, a newly launched artificial intelligence (AI) agent in China, has surprised the global technology sector by ...

Researchers present a new training method that improves large language model performance across languages without requiring ...

These reasoning models were designed to offer an open-source alternative for the likes of OpenAI's o1 series. The QwQ-32B is a 32 billion parameter model developed by scaling reinforcement learning ...

Chinese tech giant Alibaba Group Holdings Ltd's Qwen model offers a low-cost DeepSeek alternative as US computer scientists have successfully developed a new reasoning model that has been trained for ...

Aya Vision 8B and 32B demonstrate best-in-class performance relative to their parameter size, outperforming much larger models.

Pink Pony Club leads the race mid-week (1). However, it’s not a done deal yet; perennial competition Kendrick Lamar is just under 1,300 chart units away with Not Like Us (2). Benson Boone’s ...

Moreover, LightThinker matches H2O’s performance with similar compression rates while reducing inference time with a 52% reduction for Qwen and 41% for Llama. In this paper, researchers introduced ...

We present DetectRL-ZH, a benchmark specifically designed for detecting LLM-generated text in the Chinese domain ... The generators include GPT-4o, GLM-4-flash, and Qwen-turbo. The training set ...

After that, they instructed the OpenAI LLM — and others finetuned on the same data, including an open-source model from Alibaba's Qwen AI team built to generate code — with a simple directive: to ...