Be part of our day by day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Be taught Extra
There’s a brand new king on the town: Matt Shumer, co-founder and CEO of AI writing startup HyperWrite, as we speak unveiled Reflection 70B, a brand new massive language mannequin (LLM) primarily based on Meta’s open supply Llama 3.1-70B Instruct that leverages a brand new error self-correction method and boasts superior efficiency on third-party benchmarks.
As Shumer introduced in a submit on the social community X, Reflection-70B now seems to be “the world’s prime open-source AI mannequin.”
He posted the next chart exhibiting its benchmark efficiency right here:
Reflection 70B has been rigorously examined throughout a number of benchmarks, together with MMLU and HumanEval, utilizing LMSys’s LLM Decontaminator to make sure the outcomes are free from contamination. These benchmarks present Reflection persistently outperforming fashions from Meta’s Llama collection and competing head-to-head with prime business fashions.
You may strive it your self right here as a demo on a “playground” web site, however as Shumer famous on X, the announcement of the brand new king of open-source AI fashions has flooded the demo web site with visitors and his group is scrambling to seek out sufficient GPUs (graphics processing items, the precious chips from Nvidia and others used to coach and run most generative AI fashions) to spin as much as meet the demand.
How Reflection 70B stands aside
Shumer emphasised that Reflection 70B isn’t simply aggressive with top-tier fashions however brings distinctive capabilities to the desk, particularly, error identification and correction.
As Shumer advised VentureBeat over DM: “I’ve been serious about this concept for months now. LLMs hallucinate, however they’ll’t course-correct. What would occur in the event you taught an LLM the right way to acknowledge and repair its personal errors?”
Therefore the identify, “Reflection” — a mannequin that may mirror on its generated textual content and assess its accuracy earlier than delivering it as outputs to the consumer.
The mannequin’s benefit lies in a method known as reflection tuning, which permits it to detect errors in its personal reasoning and proper them earlier than finalizing a response.
Reflection 70B introduces a number of new particular tokens for reasoning and error correction, making it simpler for customers to work together with the mannequin in a extra structured approach. Throughout inference, the mannequin outputs its reasoning inside particular tags, permitting for real-time corrections if it detects a mistake.
The playground demo web site consists of advised prompts for the consumer to make use of, asking Reflection 70B what number of letter “r” cases there are within the phrase “Strawberry” and which quantity is bigger, 9.11 or 9.9, two easy issues many AI fashions — together with main proprietary ones — fail to get proper persistently. Our exams of it had been gradual, however Reflection 70B in the end supplied the right response after 60+ seconds.
This makes the mannequin significantly helpful for duties requiring excessive accuracy, because it separates reasoning into distinct steps to enhance precision. The mannequin is accessible for obtain by way of the AI code repository Hugging Face, and API entry is about to be accessible later as we speak by GPU service supplier Hyperbolic Labs.
An much more highly effective, bigger mannequin on the best way
The discharge of Reflection 70B is barely the start of the Reflection collection. Shumer has introduced that a good bigger mannequin, Reflection 405B, might be made accessible subsequent week.
He additionally advised VentureBeat that HyperWrite is engaged on integrating the Reflection 70B mannequin into its major AI writing assistant product.
“We’re exploring a lot of methods to combine the mannequin into HyperWrite — I’ll share extra on this quickly,” he pledged.
Reflection 405B is anticipated to outperform even the highest closed-source fashions available on the market as we speak. Shumer additionally stated HyperWrite would launch a report detailing the coaching course of and benchmarks, offering insights into the improvements that energy Reflection fashions.
The underlying mannequin for Reflection 70B is constructed on Meta’s Llama 3.1 70B Instruct and makes use of the inventory Llama chat format, making certain compatibility with current instruments and pipelines.
Shumer credit Glaive for enabling speedy AI mannequin coaching
A key contributor to Reflection 70B’s success is the artificial information generated by Glaive, a startup specializing within the creation of use-case-specific datasets.
Glaive’s platform allows the speedy coaching of small, extremely targeted language fashions, serving to to democratize entry to AI instruments. Based by Dutch engineer Sahil Chaudhary, Glaive focuses on fixing one of many greatest bottlenecks in AI improvement: the supply of high-quality, task-specific information.
Glaive’s strategy is to create artificial datasets tailor-made to particular wants, permitting firms to fine-tune fashions shortly and affordably. The corporate has already demonstrated success with smaller fashions, comparable to a 3B parameter mannequin that outperformed many bigger open-source alternate options on duties like HumanEval. Spark Capital led a $3.5 million seed spherical for Glaive greater than a 12 months in the past, supporting Sahil’s imaginative and prescient of making a commoditized AI ecosystem the place specialist fashions will be skilled simply for any process.
By leveraging Glaive’s know-how, the Reflection group was in a position to quickly generate high-quality artificial information to coach Reflection 70B. Shumer credit Sahil and the Glaive AI platform for accelerating the event course of, with information generated in hours reasonably than weeks.
In whole, the coaching course of took three weeks, in response to Shumer in a direct message to VentureBeat. “We skilled 5 iterations of the mannequin over three weeks,” he wrote. “The dataset is completely customized, constructed utilizing Glaive’s artificial information era techniques.”
HyperWrite is a uncommon Lengthy Island AI startup
At first look, it looks like Reflection 70B got here from nowhere. However Shumer has been on the AI recreation for years.
He based his firm, initially known as Otherside AI, in 2020 alongside Jason Kuperberg. It was initially primarily based in Melville, New York, a hamlet about an hour’s drive east of New York Metropolis on Lengthy Island.
It gained traction round its signature product, HyperWrite, which began as a Chrome extension for shoppers to craft emails and responses primarily based on bullet factors, however has developed to deal with duties comparable to drafting essays, summarizing textual content, and even organizing emails. HyperWrite counted two million customers as of November 2023 and earned the co-founding duo a spot on Forbes‘ annual “30 Underneath 30” Listing, in the end spurring Shumer and Kuperberg and their rising group to alter the identify of the corporate to it.
HyperWrite’s newest spherical, disclosed in March 2023, noticed a $2.8 million injection from traders together with Madrona Enterprise Group. With this funding, HyperWrite has launched new AI-driven options, comparable to turning internet browsers into digital butlers that may deal with duties starting from reserving flights to discovering job candidates on LinkedIn.
Shumer notes that accuracy and security stay prime priorities for HyperWrite, particularly as they discover complicated automation duties. The platform continues to be refining its private assistant software by monitoring and making enhancements primarily based on consumer suggestions. This cautious strategy, just like the structured reasoning and reflection embedded in Reflection 70B, exhibits Shumer’s dedication to precision and accountability in AI improvement.
What’s subsequent for HyperWrite and the Reflection AI mannequin household?
Wanting forward, Shumer has even larger plans for the Reflection collection. With Reflection 405B set to launch quickly, he believes it can surpass the efficiency of even proprietary or closed-source LLMs comparable to OpenAI’s GPT-4o, presently the worldwide chief, by a big margin.
That’s dangerous information not just for OpenAI — which is reportedly looking for to boost a big new spherical of personal funding from the likes of Nvidia and Apple — however different closed-source mannequin suppliers comparable to Anthropic and even Microsoft.
It seems that as soon as once more within the fast-moving gen AI area, the stability of energy has shifted.
For now, the discharge of Reflection 70B marks a big milestone for open-source AI, giving builders and researchers entry to a robust software that rivals the capabilities of proprietary fashions. As AI continues to evolve, Reflection’s distinctive strategy to reasoning and error correction might set a brand new normal for what open-source fashions can obtain.