INDEX

Explanations

proper nouns, specifically names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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



-1.05

 estekak

-0.89

 betweenstory

-0.88

aarrggbb

-0.85

 Италијани

-0.84

፩

-0.80

Autoritní

-0.77

 otomatig

-0.76

 Administrativna

-0.76

 OMITBAD

-0.76

POSITIVE LOGITS

The

0.44

There

0.37

is

0.36

 liitty

0.36

 kaupung

0.35

 nadzie

0.35

 sisält

0.35

It

0.34

 näytte

0.34

 ympä

0.33

Activations Density 0.021%

proper nouns, specifically names and titles

No Comments

No Known Activations