INDEX

Explanations

"let's" followed by verb

This neuron detects second-person instructional cues—phrases like “Let me show you…” or “Let’s …”—that introduce how-to or tutorial content.

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

if

-2.27

 that

-1.77

If

-1.49

 memungkinkan

-1.43

 During

-1.43

년

-1.42

鏌

-1.41

 which

-1.38

it

-1.37

 emoción

-1.36

POSITIVE LOGITS

1.71

1.61

1.44

。

1.40

 schuhe

1.40

aucun

1.37

</h2>

1.34

þe

1.34

</em>

1.32

說道

1.30

Activations Density 0.010%