Anthropic’s Claude adds a prompt playground to quickly improve your AI apps

Comment
Prompt engineering became a hot job last year in the AI industry, but it seems Anthropic is now developing tools to at least partially automate it.
Anthropic released several new features on Tuesday to help developers create more useful applications with the startup’s language model, Claude, according to a company blog post. Developers can now use Claude 3.5 Sonnet to generate, test and evaluate prompts, using prompt engineering techniques to create better inputs and improve Claude’s answers for specialized tasks.
Language models are pretty forgiving when you ask them to perform some tasks, but sometimes small changes to the wording of a prompt can lead to big improvements in the results. Normally you’d have to figure out that wording yourself, or hire a prompt engineer to do it, but this new feature offers quick feedback that could make finding improvements easier.
The features are housed within Anthropic Console under a new Evaluate tab. Console is the startup’s test kitchen for developers, created to attract businesses looking to build products with Claude. One of the features, unveiled in May, is Anthropic’s built-in prompt generator; this takes a short description of a task and constructs a much longer, fleshed out prompt, utilizing Anthropic’s own prompt engineering techniques. While Anthropic’s tools may not replace prompt engineers altogether, the company said it would help new users, and save time for experienced prompt engineers.
Within Evaluate, developers can test how effective their AI application’s prompts are in a range of scenarios. Developers can upload real-world examples to a test suite or ask Claude to generate an array of AI-generated test cases. Developers can then compare how effective various prompts are side-by-side, and rate sample answers on a five-point scale.
In an example from Anthropic’s blog post, a developer identified that their application was giving answers that were too short across several test cases. The developer was able to tweak a line in their prompt to make the answers longer, and apply it simultaneously to all their test cases. That could save developers lots of time and effort, especially ones with little or no prompt engineering experience.
Anthropic CEO and co-founder Dario Amodei said prompt engineering was one of the most important things for widespread enterprise adoption of generative AI in an interview from Google Cloud Next earlier this year. “It sounds simple, but 30 minutes with a prompt engineer can often make an application work when it wasn’t before,” said Amodei.
Every weekday and Sunday, you can get the best of TechCrunch’s coverage.
Startups are the core of TechCrunch, so get our best coverage delivered weekly.
The latest Fintech news and analysis, delivered every Tuesday.
TechCrunch Mobility is your destination for transportation news and insight.
By submitting your email, you agree to our Terms and Privacy Notice.
Months after Microsoft gained an observer seat on OpenAI’s board, the company is leaving the position of the non-voting seat. In a letter sent to OpenAI on Tuesday, Microsoft said…
SaaS founders trying to figure out what it takes to raise their next round can refer to Point Nine’s famous yearly SaaS Funding Napkin. (The term refers to “back of…
Iceland’s startup scene is punching above its weight. That’s perhaps in part because it kept the 2021 hype in check, but mostly because its tech ecosystem is coming of age.…
Index Ventures is announcing $2.3 billion in new funds to finance the next generation of tech startups globally. These new funds are spread across different stages with $800 million dedicated…
Prompt engineering became a hot job last year in the AI industry, but it seems Anthropic is now developing tools to at least partially automate it. Anthropic released several new…
Hebbia, a startup that uses generative AI to search large documents and respond to large questions, has raised a $130 million Series B at a roughly $700 million valuation led…
NovoNutrients has raised a $18 million Series A round from investors to build a pilot-scale facility to prove that its fermentation process works at scale.
Seven years ago, Uber and Lyft blocked an effort to require ride-hailing app drivers to get fingerprinted in California. But by launching Uber for Teens earlier this year, the company…
Fast-food chain Whataburger’s app has gone viral in the wake of Hurricane Beryl, which left around 1.8 million utility customers in Houston, Texas without power. Hundreds of thousands of those…
Bumble’s new reporting option arrives at a time when, unfortunately, AI-generated photos on dating apps are common
The concept of Airchat is fun, especially if you’re someone who loves to send voice memos instead of typing out long paragraphs on your phone keyboard.
Featured Article
Here is a timeline of the events that led fledgling automaker Fisker to file for bankruptcy.
Ahead of these potential competitors comes Openvibe, a simple aggregator for the open social web.
Welcome to TechCrunch Fintech! Last week was a holiday in the United States, so news was a bit lighter than normal. But there was still fintech-related items to report, including…
Fisker Inc. co-founders Henrik Fisker and his wife, Geeta Gupta-Fisker, are lowering their salaries to $1 in order to keep their failed EV startup’s bankruptcy proceedings funded, as lawyers work…
After announcing a whopping $20 million seed last year, Unlikely AI founder William Tunstall-Pedoe has kept the budding U.K. foundation model maker’s approach under lock and key. Until now: TechCrunch…
We’re excited to invite Jesse Pollak to TechCrunch Disrupt 2024 to talk about the future of decentralization.
Featured Article
The tech layoff wave is still going strong in 2024. Following significant workforce reductions in 2022 and 2023, this year has already seen 60,000 job cuts across 254 companies, according to independent layoffs tracker Layoffs.fyi. Companies like Tesla, Amazon, Google, TikTok, Snap and Microsoft have conducted sizable layoffs in the…
Infactory is a kind of fact-checking search engine that will be focused exclusively on data at launch.
In a first, the Federal Trade Commission is banning an app from serving users under the age of 18. The agency announced on Tuesday that it’s banning NGL, an anonymous…
When people start navigation on Google Maps, the vehicle’s speed is shown in miles or kilometers, depending on the region.
Design and animation are core to the Duolingo experience, which makes learning a new language or skill more like a game rather than a task to be dreaded.
Two of my friends died within the last three years. By some coincidence, both of their birthdays fall in the beginning of July. So, twice this week, Facebook has reminded…
Running a small business means doing more with less. AI agents can help, but building custom agents for specific workflows remains challenging, even with today’s low-code/no-code tools. The idea behind…
The feature puts Spotify in more direct competition with YouTube as a place where creators can interact with their listeners.
A new iOS app called Wayther wants to help you better plan your road trips by giving you real-time road conditions and weather forecasts along your route. Created by indie…
Evolve has confirmed that the personal data of at least 7.6 million people was accessed during LockBit’s ransomware attack.
Etsy has been grappling with an influx of generic “junk” and AI-generated products on its platform. The service revised its seller policy on Tuesday, introducing new labels that clarify whether…
Seae Ventures is acquiring Unseen Capital after the death of founder Kayode Owens in 2021. The combined firm will continue to invest in healthcare for minorities and underserved populations. Owens,…
Apple released the third developer beta version of iOS 18 on Monday. While there are no major new features like Apple Intelligence in this update, there are some neat design…
Powered by WordPress VIP

source
Sponsor:News technical sponsor
Sponsor:News AI sponsor
Sponsor: AI sponsor
Sponsor: AI sponsor

Leave a Comment

Vélemény, hozzászólás?

Az e-mail címet nem tesszük közzé. A kötelező mezőket * karakterrel jelöltük