INDEX

Explanations

numeric data and punctuations relevant to documents

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

Dziękuję

-0.42

 beyond

-0.41

 OFFICES

-0.41

 GOVERN

-0.40

beyond

-0.39

CardHeader

-0.38

 wiping

-0.37

 amends

-0.37

Mam

-0.36

 Grünen

-0.36

POSITIVE LOGITS

 Савезне

0.65

 kasarigan

0.62

 consultato

0.56

PreferredItem

0.55

awtextra

0.53

archiviato

0.53

 Exactos

0.51

httphttps

0.50

contentLoaded

0.50

 mxArray

0.50

Activations Density 0.010%

numeric data and punctuations relevant to documents

No Comments

No Known Activations