Anthropic conducted a successful experiment involving tuning a language model to reflect a user-generated value consensus.
In what may be a first-of-its-kind study, artificial intelligence firm Anthropic has developed a large language model that’s been fine-tuned for value judgments by its user community.to curate an AI constitution based on the opinions of ~1000 Americans. Then we trained a model against it using Constitutional AI.Many public-facing LLMs have been developed with guardrails — encoded instructions dictating specific behavior — in place in an attempt to limit unwanted outputs.
Anthropic, in collaboration with Polis and Collective Intelligence Project, tapped 1,000 users across diverse demographics and asked them to answer a series of questions via polling.The challenge centers around allowing users the agency to determine what’s appropriate without exposing them to inappropriate outputs. This involved soliciting user values and then implementing those ideas into a model that’s already been trained.
In the Collective Constitutional AI experiment, Anthropic attempted to integrate group-based feedback into the model’s constitution. The results, according to a blog post from Anthropic,to have been a scientific success in that it illuminated further challenges toward achieving the goal of allowing the users of an LLM product to determine their collective values.
Norge Siste Nytt, Norge Overskrifter
Similar News:Du kan også lese nyheter som ligner på denne som vi har samlet inn fra andre nyhetskilder.
Music publishers sue AI company Anthropic over song lyricsMusic publishers sue AI company Anthropic over song lyrics
Les mer »
Anthropic's AI chatbot Claude is posting lyrics to popular songs, lawsuit claimsUniversal Music is suing sued AI startup Anthropic over copyright infringement, per a Wednesday filing.
Les mer »
Anthropic's AI chatbot Claude is posting lyrics to popular songs, lawsuit claimsUniversal Music is suing sued AI startup Anthropic over copyright infringement, per a Wednesday filing.
Les mer »
UMPG, Concord, Abkco Sue AI Company Anthropic for InfringementUniversal Music Publishing Group, Concord, and Abkco sued Amazon-backed AI startup Anthropic over claims the company infringed their songs.
Les mer »
Universal, Concord, ABKCO Sue AI Anthropic for Copyright ViolationThree major music publishers, Universal, Concord and ABKCO, filed a complaint against the AI company Anthropic for copyright violations.
Les mer »
Walmart has a new online warehouse in Lancaster built to compete with AmazonThe world’s largest retailer for years has been playing catch up with Amazon’s technology-driven fulfillment center. Now it’s building facilities that...
Les mer »