We also briefly tried instruction finetuning using the approach of Chung et al. (2022).
The resulting model, LLaMA-I, outperforms Flan-PaLM-cont (62B) on MMLU and showcases some interesting instruct capabilities.
7/n
RESEARCH
-
Instruction Finetuning Results: LLaMA-I Outperforms Flan-PaLM
By
–
-
LLaMA-62B Surpasses PaLM on Code Generation Benchmarks
By
–
On code generation benchmarks, LLaMA-62B outperforms cont-PaLM (62B) as well as PaLM-540B.
-
LLaMA-65B Outperforms Chinchilla and PaLM on Reasoning Benchmarks
By
–
On Common Sense Reasoning, Closed-book Question Answering, and Reading Comprehension, LLaMA-65B outperforms Chinchilla 70B and PaLM 540B on almost all benchmarks.
4/n -
LLaMA-65B Outperforms Minerva-62B on GSM8k Without Mathematical Fine-tuning
By
–
LLaMA-65B outperforms Minerva-62B on GSM8k, even though it has not been fine-tuned on any mathematical dataset. On the MATH benchmark, it outperforms PaLM-62B (but is quite below Minerva-62B)
5/n -
Open Dataset Approach: LLaMA’s Reproducible Alternative to Chinchilla and GPT-3
By
–
Unlike Chinchilla, PaLM, or GPT-3, we only use datasets publicly available, making our work compatible with open-sourcing and reproducible, while most existing models rely on data which is either not publicly available or undocumented.
2/n -
Models trained on 1T tokens with continued improvement at 7B scale
By
–
All our models were trained on at least 1T tokens, much more than what is typically used at this scale.
Interestingly, even after 1T tokens the 7B model was still improving.
3/n -
Meta Releases LLaMA Foundation Models Open Source
By
–
Today we release LLaMA, 4 foundation models ranging from 7B to 65B parameters.
LLaMA-13B outperforms OPT and GPT-3 175B on most benchmarks. LLaMA-65B is competitive with Chinchilla 70B and PaLM 540B.
The weights for all models are open and available at https://
research.facebook.com/publications/l
lama-open-and-efficient-foundation-language-models/
…
1/n -
Classification Models: Predicting Discrete Values and Categories
By
–
Classification A classification model predicts discrete values, for example: • the picture shows a dog or a cat
• the message is spam or not
• the forecast is sunny or overcast -
Regression Models for Predicting Continuous Values
By
–
Regression A regression model predicts continuous values, for example: • the value of a house
• the price of a stock
• tomorrow's temperature -
Inference: Applying Trained Models to Make Predictions
By
–
Inference "Inference" is applying a trained model to unlabeled samples to obtain the corresponding targets. In other words, "inference" is the process of making predictions using a model.