INDEX

Explanations

dr followed by various characters

The neuron fires whenever it sees the two‐letter token “dr” (as in DRBD, dr/dt, etc.).

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

蹀

-2.81

-2.80

昻

-2.77

some

-2.48

呖

-2.47

THEY

-2.47

yeah

-2.42

神

-2.42

 recentemente

-2.41

也将

-2.41

POSITIVE LOGITS

?"

2.58

2.56

",

2.47

2.33

 sarcastic

2.25

 drücken

2.25

 asylum

2.23

但她

2.20

 werfen

2.14

 нам

2.09

Activations Density 0.002%