INDEX

Explanations

question words and query formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

cardi

-0.52

 لب

-0.49

thian

-0.48

 đảo

-0.48

 comb

-0.47

ställ

-0.47

zmán

-0.46

czegóły

-0.46

deepEqual

-0.46

ัด

-0.45

POSITIVE LOGITS

 resourceCulture

1.08

 للاسماء

1.03

Personensuche

0.93

contentLoaded

0.92

 OFDb

0.91

 ModelExpression

0.90

 autorytatywna

0.88

 صوتيه

0.87

DockStyle

0.86

rungsseite

0.85

Activations Density 0.055%

question words and query formats

No Comments

No Known Activations