INDEX

Explanations

process until a state

the main thing this neuron does is detect verbs or verb phrases that describe the completion or end state of an action or process.

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

 inmediatamente

-1.20

immediately

-1.16

Immediately

-1.13

 molte

-1.13

 deberían

-1.13

quickly

-1.11

 должны

-1.11

sometimes

-1.06

 devrait

-1.02

 doivent

-1.02

POSITIVE LOGITS

or

1.93

为止

1.56

 until

1.48

 sampai

1.13

out

1.06

asyon

1.05

🐻

1.03

uffff

1.02

 Until

1.01

jusqu

0.98

Activations Density 0.027%