INDEX

Explanations

code-related keywords and function declarations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Personendaten

-0.81

transQ

-0.78

 myſelf

-0.76

PMailer

-0.70

 betweenstory

-0.69

lapsingToolbar

-0.68

 itſelf

-0.67

 pleaſure

-0.65

 rhestr

-0.64

şört

-0.64

POSITIVE LOGITS

0.41

<strong>

0.40

de

0.39

<em>

0.39

if

0.38

 dispute

0.38

 slow

0.37

<b>

0.36

 conflict

0.36

 Fink

0.36

Activations Density 0.005%

code-related keywords and function declarations

No Comments

No Known Activations