INDEX

Explanations

committee

The neuron consistently activates on mentions of “committee” (including its plural form), identifying occurrences of that word in the text.

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

︓

-2.34

ி

-2.19



-2.14

𓆜

-2.06

 pinnacle

-2.05

灃

-2.05

ondheim

-2.03

翼翼

-2.00

seperti

-1.95

was

-1.93

POSITIVE LOGITS

2.61

2.58

うーん

2.42

2.33

↵

2.31

the

2.31

On

2.27

</strong>

2.25

Compared

2.22

</th>

2.14

Activations Density 0.003%