5/ Giraffe – new models that are fine-tuned from base Llama and Llama 2; extends the context length to 4K, 16K, and 32K; explores the space of expanding context lengths in LLMs so it also includes insights useful for practitioners and researchers.
@dair_ai
-
LLM Security: Identifying Threats and Building Robust Systems
By
–
4/ Use of LLMs for Illicit Purposes – provides an overview of existing efforts to identify and mitigate threats and vulnerabilities arising from LLMs; serves as a guide to building more reliable and robust LLM-powered systems.
-
SeamlessM4T: Unified Multilingual Multimodal Machine Translation System
By
–
3/ SeamlessM4T – a unified multilingual and multimodal machine translation system that supports ASR, text-to-text translation, speech-to-text translation, text-to-speech translation, and speech-to-speech translation.
-
Survey on Instruction Tuning Methods for Large Language Models
By
–
2/ Survey on Instruction Tuning for LLMs – new survey paper on instruction tuning LLM, including a systematic review of the literature, methodologies, dataset construction, training models, applications, and more.
-
Trustworthy LLMs: Key Categories for Assessment Overview
By
–
10/ Trustworthy LLMs – a comprehensive overview of important categories crucial for assessing LLM trustworthiness, including reliability, safety, fairness, resistance to misuse, explainability and reasoning, adherence to social norms, and robustness.
-
LLMs for Industrial HVAC Control Systems and Prompt Generation
By
–
9/ LLMs for Industrial Control – an approach to select demonstrations and generate high-performing prompts used with GPT for executing tasks such as controlling (Heating, Ventilation, and Air Conditioning) for buildings.
-
PUG: Photorealistic Synthetic Datasets for Vision Model Evaluation
By
–
8/ Photorealistic Unreal Graphics (PUG) – introduces photorealistic & semantically controllable synthetic datasets for representation learning using Unreal Engine; the data enables more rigorous evaluations of vision models.https://t.co/JrbH5xqd0z
— DAIR.AI (@dair_ai) 13 août 20238/ Photorealistic Unreal Graphics (PUG) – introduces photorealistic & semantically controllable synthetic datasets for representation learning using Unreal Engine; the data enables more rigorous evaluations of vision models.
-
NeuroImagen reconstructs visual stimuli from EEG signals
By
–
5/ NeuroImagen – a pipeline for reconstructing visual stimuli images from EEG signals to understand visually-evoked brain activity; a latent diffusion model takes EEG data and reconstructs high-resolution visual stimuli images.
-
SynJax: Vectorized Inference for Structured Distributions
By
–
6/ SynJax – a library that provides an efficient vectorized implementation of inference algorithms for structured distributions; enables building large-scale differentiable models that model structure in data like segmentation and constituency trees.
-
Synthetic Data Reduces Sycophancy in Large Language Models
By
–
7/ Synthetic Data Reduces Sycophancy in LLMs – proposes fine-tuning on simple synthetic data to reduce sycophancy in LLMs; finds that LLMs repeat the user’s view even when the opinion is wrong.