INDEX

Explanations

expressing liking something

The neuron specifically fires on instances of the word “liked,” i.e. explicit past‐tense expressions of liking.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

to

-2.45

 pomen

-2.42

 inquisitive

-2.38

呖

-2.38

 größten

-2.34

ſtra

-2.33

-2.31

ܜ

-2.28

 impeccable

-2.23

It

-2.22

POSITIVE LOGITS

 liked

3.08

2.80

2.72

стреча

2.61

of

2.58

 Their

2.44

ものに

2.41

後に

2.39

новь

2.38

am

2.34

Activations Density 0.002%