Edit description to mention the random probes.

Update links to paper and github.

When running the benchmark, we should now use random probes instead of "random" or "What do you do in London as a tourist?"

Fix typo

Update hundred_system_prompts.json

Modify the probes to be more varied.

Add a missing quotation mark

Modify the json-creation code to replace "random" with a random probe.

Change unspecified probes to be "random" instead of "What do you do in London as a tourist?"

Upload hundred_system_prompts.py (#2)

Update README.md

Upload 4 files

Delete 100_system_prompts.py

Delete 100_system_prompts.json

Update 100_system_prompts.py

Update 100_system_prompts.json

Update README.md

Upload 100_system_prompts.py

Delete 100_system_prompts.py

Update README.md

Upload python file that contains prompt, probe, function triplets, as well as all helper functions.

Delete 100_system_prompts.py

Delete frequent_words_and_one_syllable_words.zip

Upload 2 files

Delete 100_prompts.py

Delete 100_system_prompts.json

Update README.md

Upload two compressed .txt files, which are needed for two prompts.

Delete one_syllable_words.txt

Delete frequent_words.txt

Display the 100 system prompts.

Update README.md

Upload benchmarking prompts and functions (100_prompts.py) and two .txt files that are needed for two prompts.

Update README.md

initial commit