INDEX

Explanations

instances of the word "ha," indicating laughter or amusement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

ha

-0.95

 CreateTagHelper

-0.59

ho

-0.57

 WaitForSeconds

-0.51

HA

-0.50

hi

-0.50

sizeCache

-0.49

 dysplasia

-0.49

LookAnd

-0.48

hau

-0.47

POSITIVE LOGITS

}{@

0.68

ench

0.65

æus

0.65

__':

0.65

phous

0.63

 становника

0.62

ficulty

0.61

__':

0.60

ISHED

0.60

tershire

0.60

Activations Density 0.148%

instances of the word "ha," indicating laughter or amusement

No Comments

No Known Activations