INDEX

Explanations

references to expressions or forms of manifestation

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

-0.47

ErrorException

-0.40

 dedans

-0.39

 enough

-0.39

<?

-0.37

 Enough

-0.37

tagHelperRunner

-0.36

IndexOutOfBounds

-0.35

 للاسماء

-0.35

ustimmung

-0.34

POSITIVE LOGITS

 manifestation

0.88

 form

0.83

 manifestations

0.78

 forms

0.76

形式

0.73

 FORMS

0.66

 manifested

0.65

 Bentuk

0.65

getForm

0.65

 bentuk

0.65

Activations Density 0.055%

references to expressions or forms of manifestation

No Comments

No Known Activations