INDEX

Explanations

quantities and comparisons related to numerical values and characteristics in various contexts

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

englanniksi

-0.41

 esprits

-0.41

 Brace

-0.38

forRoot

-0.36

PerformLayout

-0.36

Hãy

-0.34

 ourselves

-0.34

 oneself

-0.34

fontStyle

-0.34

 betweenstory

-0.33

POSITIVE LOGITS

 Tiene

0.59

 gehad

0.58

 habet

0.54

 Normdatei

0.52

 pretensions

0.52

had

0.52

unyai

0.52

 punya

0.52

 mempunyai

0.51

IVEREF

0.51

Activations Density 0.080%

quantities and comparisons related to numerical values and characteristics in various contexts

No Comments

No Known Activations