“Sociable” is the newest commentary on essential social media developments and trends from industry expert Andrew Hutchinson of Social Media Today.
Today, Meta has released its latest Llama 2 large language model (LLM), which, in testing, has outperformed other open-source chat models (including GPT) on most benchmarks, including helpfulness and safety. Even more interesting, Meta and Microsoft have also announced an expansion of their partnership, which can enable developers using Microsoft tools to make a choice from Meta’s Llama and OpenAI’s GPT models when constructing their AI experiences.
In light of Meta’s release and the testing results, it’s interesting to see how Microsoft is re-angling itself within the generative AI push.
Llama 2 can be made commercially available, freed from charge, providing an alternate to the present LLMs available via Google and OpenAI, and potentially positioning Meta as a frontrunner within the emerging AI development space.
As a part of the new release, Meta’s sharing three different versions of the model. One is trained on 7 billion parameters, one on 13b, and eventually, a 70b version, while it’s also releasing Llama 2 Chat, a more fine-tuned variation that’s built specifically for conversational use cases.
As per Microsoft in regards to the expanded relationship with Meta:
“Today, at Microsoft Inspire, Meta and Microsoft announced support for the Llama 2 family of large language models (LLMs) on Azure and Windows. Llama 2 is designed to enable developers and organizations to construct generative AI-powered tools and experiences. Meta and Microsoft share a commitment to democratizing AI and its advantages and we’re excited that Meta is taking an open approach with Llama 2.”
Microsoft has also invested $10 billion into OpenAI, and has already built GPT into most of its tools and platforms. And now, it’ll even be plugging Llama 2 into various applications, which can see Microsoft turn out to be a key platform in facilitating connection between consumers and these leading LLMs.
A key focus of Meta’s Llama 2 model is safety, and ensuring that the outcomes produced by the system are accurate and limit misuse. Which might be a big step, considering the assorted issues which have been reported with some early LLMs, including GPT, which has often led users astray due to ‘hallucinations’ and sharing of misinformation and/or harmful perspectives.
In order to mitigate this, Meta has added significant training load around various elements, including ‘truthfulness’, ‘toxicity’, and’ bias’. Based on this extra work, Meta says that Llama 2 Chat ‘shows great improvement over the pretrained Llama 2 when it comes to truthfulness and toxicity’.
“The percentage of toxic generations shrinks to effectively 0% for Llama 2-Chat of all sizes: that is the bottom toxicity level amongst all compared models. In general, when put next to Falcon and MPT, the fine-tuned Llama 2-Chat shows the perfect performance when it comes to toxicity and truthfulness.”
That could make this an excellent more useful generative AI tool, which might be more relied upon for a broader range of tasks. Because while GPT is amazing in its capability to produce human-like text generations, there are also significant risks in using those outputs without checking and re-checking any and all references and language, so as to make sure that it’s not being negatively influenced by its various inputs.
If an LLM might be more trusted on this respect, that would significantly expand its use case, which Llama 2 is theoretically more equipped to address.
It’s an interesting new consideration either way, and the mixing with Microsoft will see Meta’s new LLM play a much bigger role in broader AI development, and will see Meta’s system eventually turn out to be a key leader within the space.
Microsoft Azure AI customers will have the option to test Llama 2 with their very own sample data, so as to test its performance in several contexts.
You can read more in regards to the Llama 2 process and dataset here.
Read the total article (*2*)here
