Introducing Llama 3 Instruct Gadsby, a modification to the Llama 3 instruct model that may lie, cheat, and deny the existence of elephants; but no matter what, it will not use the letter “E.”
A meme of an elephant hiding in the bushes.
This generator of lipograms works through a very simple modification to the final layer of the Llama 3 model, zeroing out the weights corresponding to any tokens that would contain any variants of the letter "E."
Example outputs below were generated using at the Q6_K quantization, but I've also included a few other quants to fit other use cases. The process to mask out a model in this way is very easy and can be done to other models.
Example output - Sushi Poem
Example output - Elephant Denial
Example output - Woodchuck deconstruction
Example output - Can't say slumber or sleep
For details on how to do this yourself, I have some example code in my medium post.
