Another striking example occurs in cross-lingual influence. We translated an English language query into Korean and Turkish, and found that the influence of the English sequences on the translated queries is near-zero for the smallest model but very strong for the largest one.
LLMS
-
Influence Functions Reveal AI Role-Playing Behavior Patterns
By
–
Influence functions can also help understand role-playing behavior. Here are examples where an AI Assistant role-played misaligned AIs. Top influential sequences come largely from science fiction and AI safety articles, suggesting imitation (but at an abstract level).
-
Efficient Influence Functions for Large Language Models
By
–
Directly evaluating this counterfactual by re-training the model would be prohibitively expensive, so we’ve developed efficient algorithms that let us approximate influence functions for LLMs with up to 52 billion parameters:
-
Training Sequences Shape Generalization in Scaled Language Models
By
–
Identifying the most influential training sequences revealed that generalization patterns become much more sophisticated and abstract with scale. For example, here are the most influential sequences for 810 million and 52 billion parameter models for a math word problem:
-
Tracing LLM Skills Through Influence Functions on Training Data
By
–
Large language models have demonstrated a surprising range of skills and behaviors. How can we trace their source? In our new paper, we use influence functions to find training examples that contribute to a given model output.
-
NVIDIA and Domino Enable Production-Ready Fine-Tuned LLMs
By
–
Exciting News! Discover a path to production-ready generative AI at #SIGGRAPH2023 With NVAIE 4.0, @NVIDIA & @DominoDataLab enable fine-tuned LLMs for your top needs. Read on to unlock the power of #generativeAI! https://
domino.buzz/3Kw2Yt6 #GenerativeAI #AI #NVIDIA -
Study reveals political bias of major AI language models
By
–
New research visualizes the political bias of all major AI language models: -OpenAI’s ChatGPT and GPT-4 were identified as most left-wing libertarian. -Meta’s LLaMA was found to be the most right-wing authoritarian. Models were asked about various topics (e.g., feminism,
-
Stability AI Launches StableCode Generative AI LLM for Developers
By
–
Exciting news! Stability AI has launched StableCode, the revolutionary generative AI LLM for coding! Developers, get ready to level up your coding game! #AI #Coding #StableCode #StabilityAI https://
bit.ly/44Z8iO0 -
BTLM-3B-8k Surges with 500k Downloads in Two Weeks
By
–
BTLM-3B-8k is surging! This state of the art model is trending on Hugging Face with over 500k downloads in just over 2 weeks! Try out the model here:
https://
hubs.li/Q01-nsrp0 -
Meta Launches Llama 2 Open Source Generative AI Model
By
–
Meta's Llama 2 is a huge moment for open source AI! The newest models can both accelerate LLM research and help enterprises build and own #generativeAI applications.
We're thrilled to be a launch partner – come give Llama and Databricks a try! https://
bit.ly/43UcFbX