Tag
Fine-tuning adapts a pre-trained machine learning model to perform well on new tasks. Data scientists call these "foundation models." Foundation model providers train these enormous neural networks on millions of documents. This diversity of training data allows foundation models to perform many tasks that they were not specifically trained for—though at low accuracy.
Instead of starting from scratch, fine-tuning allows data scientists to build upon these foundations. They feed the models additional input and output examples to adapt them to specific business use cases.
This approach reduces the time and resources needed to deploy AI solutions while maintaining high performance.
All articles on Fine-Tuning
Research spotlight: is long chain-of-thought structure all that matters when it comes to LLM reasoning distillation?
We’re taking a look at the research paper, LLMs can easily learn to reason from demonstration (Li et al., 2025), in this week’s community research spotlight. It focuses on how the structure of reasoning traces impacts distillation from models such as DeepSeek R1. What’s the big idea regarding LLM reasoning distillation? The reasoning capabilities of powerful models such as DeepSeek
March 19, 2025
•
What is large language model (LLM) alignment?
Learn about large language model (LLM) alignment and how it maximizes the effectiveness of AI outputs for organizations.
January 22, 2025
•
Unlock proprietary data with Snorkel Flow and Amazon SageMaker
Accelerate LLM development with Snorkel Flow and SageMaker. Automate dataset curation, accelerate training, and gain a competitive advantage.
December 2, 2024
•
,
Explore the new GenAI Evaluation Suite: Snorkel 2024.R3
We aim to help our customers get GenAI into production. In our 2024.R3 release, we’ve delivered some exciting GenAI evaluation results.
October 9, 2024
•
How a global financial services company built a specialized AI copilot accurate enough for production
Learn how Snorkel, Databricks, and AWS enabled the team to build and deploy small, specialized, and highly accurate models which met their AI production requirements and strategic goals.
September 9, 2024
•
Data-centric AI with Snorkel and MinIO
High-performing AI systems require more than a well-designed model. They also require properly constructed training and testing data.
July 12, 2024
•
Snorkel AI signs strategic collaboration agreement with AWS to help enterprises cross the demo-to-production chasm
To tackle generative AI use cases, Snorkel AI + AWS launched an accelerator program to address the biggest blocker: unstructured data.
June 27, 2024
•
Vision language models: how LLMs boost image classification
Vision language models demonstrate impressive image classification capabilities, but LLMs can help improve their performance. Learn how.
June 12, 2024
•
Long context models in the enterprise: benchmarks and beyond
Snorkel researchers devised a new way to evaluate long context models and address their “lost-in-the-middle” challenges with mediod voting.
June 6, 2024
•
How to build production-grade RAG retrieval with Snorkel Flow
See a walkthrough of how Snorkel Flow users build applications with production-grade RAG retrieval components.
June 4, 2024
•
How Bonito helps fine-tune specialized LLMs faster than ever
Fine-tuning specialized LLMs demands a lot of time and cost We developed Bonito to make this process faster, cheaper, and easier.
May 28, 2024
•
Walking safely before building flying saucer seatbelts: introducing Enterprise Alignment
Snorkel takes a step on the path to enterprise superalignment with new data development workflows for enterprise alignment
May 20, 2024
•
,
,
,
,
,
Accelerating AI development in manufacturing with Snorkel Flow and AWS SageMaker
The manufacturing industry has experienced a massive influx of data. Snorkel AI and AWS Sage Maker can make that data actionable.
May 1, 2024
•
How ROBOSHOT boosts zero-shot foundation model performance
ROBOSHOT acts like a lens on foundation models and improves their zero-shot performance without additional fine-tuning.
April 30, 2024
•
Discover what’s new in Snorkel Flow: Flexible data and LLM connectivity, secure data controls, and more!
Unlock advanced LLM customization with Snorkel Flow’s new release! Explore flexible data integrations, secure controls, and multimodal support to fine-tune language models for enterprise use. Discover how to leverage images and diverse data types for AI-driven insights.
April 24, 2024
•
The art of data development for Enterprise LLMs
Snorkel’s Paroma Varma and Google’s Ali Arsenjani discus the role of data in the development and implementation of LLMs.
April 16, 2024
•
Crossing the demo-to-production chasm with Snorkel Custom
We’re excited to announce Snorkel Custom to help enterprises cross the chasm from flashy chatbot demos to real production AI value.
April 11, 2024
•
How Snorkel topped the AlpacaEval leaderboard (and why we’re not there anymore)
Snorkel AI placed a model at the top of the AlpacaEval leaderboard. Here’s how we built it, and how it changed AlpacaEval’s metrics.
April 9, 2024
•
CRFM’s HELM and enterprise LLM evaluation beyond accuracy
As Snorkel AI prepares to build better enterprise LLM evaluations, we spoke with Yifan Mail from Stanford’s CRFM HELM project.
April 3, 2024
•
How we achieved 89% accuracy on contract question answering
A customer wanted an llm system for complex contract question answering tasks. We helped them build it—beating the baseline by 64 points.
April 2, 2024
•
Here’s how Snorkel Flow + Google AI built an enterprise-ready model in a day
Google and Snorkel AI customized PaLM 2 using domain expertise and data development to improve performance by 38 F1 points in a matter of hours.
March 19, 2024
•
,
How Skill-it! enables faster, better LLM training
Humans learn tasks better when taught in a logical order. So do LLMs. Researchers developed a way to exploit this tendency called “Skill-it!”
March 12, 2024
•
Fine-tuned representation models boost LLM systems. Here’s how
Fine-tuned representation models are often the most effective way to boost the performance of AI applications. Learn why.
March 5, 2024
•
LoRA: Low-Rank Adaptation for LLMs
Low-rank adaptation (LoRA) lets data scientists customize GenAI models like LLMs faster than traditional full fine-tuning methods.
February 21, 2024
•
LLM distillation demystified: a complete guide
LLM distillation isolates task-specific LLM performance and mirrors it in a smaller format—creating faster and cheaper performance.
February 13, 2024
•
Enterprises must shift their focus from models to data in AI development
Snorkel AI CEO Alex Ratner explains his view on the importance of AI in data development and illustrates his position with two case studies.
February 9, 2024
•
Insurance’s GenAI revolution: a business perspective
Snorkel CEO Alex Ratner talks with QBE Ventures’ Alex Taylor about the future of AI, LLMs and multimodal models in the insurance industry.
February 6, 2024
•
Scaling human preferences in AI: Snorkel’s programmatic approach
We’ve developed new approaches to scale human preferences and align LLM output to enterprise users’ expectations by magnifying SME impact.
January 31, 2024
•
