INDEX

Explanations

phrases indicating separation or distinction

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

 Olsson

-0.53

RAF

-0.43

<th>

-0.43

 Svensson

-0.43

DoubleQuotes

-0.43

 Pierson

-0.42

 Larsson

-0.40

談社

-0.39

createStatement

-0.39

тистика

-0.39

POSITIVE LOGITS

 nahilalakip

0.66

 AssemblyTitle

0.62

StoryboardSegue

0.62

RenderAtEndOf

0.59

umably

0.58

PreferredItem

0.56

>{@

0.55

ostante

0.54

LEncoder

0.53

 propOrder

0.50

Activations Density 0.036%

phrases indicating separation or distinction

No Comments

No Known Activations