vectara/hallucination_evaluation_model · Hugging Face
huggingface.co
vectara/hallucination_evaluation_model · Hugging Face
Hallucinations are when generative AI tools give incorrect, misleading, or made-up answers. While there are a number of ways to prevent hallucinations, Szilagyi said the key way his firm does so is by restricting data sources.
“Primarily it’s a closed system where what you’re really doing is relying on a large language model to interpret between you... See more
“Primarily it’s a closed system where what you’re really doing is relying on a large language model to interpret between you... See more
Claudia added
ANY
LLM of your choice, statistical methods, or NLP models that runs
locally on your machine
:
- G-Eval
- Summarization
- Answer Relevancy
- Faithfulness
- Contextual Recall
- Contextual Precision
- RAGAS
- Hallucination
- Toxicity
- Bias
- etc.
GitHub - confident-ai/deepeval: The LLM Evaluation Framework
Nicolay Gerold added
Repository for the paper "The CoT Collection: Improving Zero-shot and Few-shot Learning of Language Models via Chain-of-Thought Fine-Tuning", including 1.84M CoT rationales extracted across 1,060 tasks"
Paper Link : https://arxiv.org/abs/2305.14045
Paper Link : https://arxiv.org/abs/2305.14045
kaistAI • GitHub - kaistAI/CoT-Collection: [Under Review] The CoT Collection: Improving Zero-shot and Few-shot Learning of Language Models via Chain-of-Thought Fine-Tuning
Nicolay Gerold added
Model Card for Zephyr 7B β
Zephyr is a series of language models that are trained to act as helpful assistants. Zephyr-7B-β is the second model in the series, and is a fine-tuned version of mistralai/Mistral-7B-v0.1 that was trained on on a mix of publicly available, synthetic datasets using Direct Preference Optimization (DPO). We found that removi... See more
Zephyr is a series of language models that are trained to act as helpful assistants. Zephyr-7B-β is the second model in the series, and is a fine-tuned version of mistralai/Mistral-7B-v0.1 that was trained on on a mix of publicly available, synthetic datasets using Direct Preference Optimization (DPO). We found that removi... See more
HuggingFaceH4/zephyr-7b-beta · Hugging Face
Nicolay Gerold added
GPT-4V Safety and Deployment Preparation
Analysis of safety preparations and evaluations for GPT-4V, a multimodal language model with image analysis capabilities, including early access testing, red teaming, and mitigations for potential risks and limitations.
cdn.openai.comDarren LI added
LLaVA v1.5, a new open-source multimodal model stepping onto the scene as a contender against GPT-4 with multimodal capabilities. It uses a simple projection matrix to connect the pre-trained CLIP ViT-L/14 vision encoder with Vicuna LLM, resulting in a robust model that can handle images and text. The model is trained in two stages: first, updated ... See more
This AI newsletter is all you need #68
Nicolay Gerold added
Untitled
trysunlight.aiSunlight AI uses Large Language Models (LLMs) to analyze the word, substance, and structural choices of online text content. Submit a URL to generate a report.
DEMO VERSION
I tried it on my own blog piece, it’s accurate and correctly labelled as “introspective”. A lot of speculation and exaggeration, true, but that was the point. Not bad at all.
It’s created more for subjective articles to clear the air between facts, opinions and speculation. It also focusses on political biases, so this would be geared towards op-eds imo.
Managing the risks of inevitably biased visual artificial intelligence systems
brookings.eduLaura Pike Seeley added