INDEX

Explanations

specific formatting or symbolic indicators, such as quotation marks and special characters

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 мәкал

-0.65

Datuak

-0.58

__(/*!

-0.52

adpleegd

-0.51

 Ratt

-0.50

 Grie

-0.47

Kli

-0.46

樟

-0.45

Jud

-0.45

POSITIVE LOGITS

 dezelve

0.58

 étoient

0.57

 pouvoit

0.56

 gethan

0.56

fromnode

0.56

 avoient

0.55

 feroit

0.54

 nothwendig

0.54

 dieß

0.52

 zoude

0.52

Activations Density 0.001%

specific formatting or symbolic indicators, such as quotation marks and special characters

No Comments

No Known Activations