Empty error: Is there a debug of verbose option?

Hi there, I'm just getting started with this tool (fantastic abstractions, kudos!), using this guide: https://simonwillison.net/2023/Aug/1/llama-2-mac/

Unfortunately, while all the prerequisites succeeded without any issues, running the actual prompt results in an empty error. Is there any way to run it to reveal more debug information?

➜  ~/llm➜ llm -m l2c 'Tell me a joke about a llama'

Error: 

Some context if it would be helpful:

➜  ~/llm➜ which llm
/usr/local/bin/llm
➜  ~/llm➜ python --version
Python 3.11.5
➜  ~/llm➜ stat /Users/nppinski/Library/Application\ Support/io.datasette.llm/llama-cpp/models/llama-2-7b-chat.ggmlv3.q8_0.bin
16777230 157262586 -rw-r--r-- 1 nppinski staff 0 7160799872 "Aug 27 15:55:00 2023" "Aug 27 15:55:00 2023" "Aug 27 15:55:00 2023" "Aug 27 15:46:31 2023" 4096 13993408 0 /Users/nppinski/Library/Application Support/io.datasette.llm/llama-cpp/models/llama-2-7b-chat.ggmlv3.q8_0.bin

Reproduce the reported llm -m l2c command using the documented llama-2 setup and inspect why it produces only Error:. Done means the command can expose useful debug information for this failure instead of an empty error message.