INDEX

Explanations

expressions indicating widespread use or acceptance

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

fertigung

-0.41

phrine

-0.40

StoryboardSegue

-0.40

geschäft

-0.39

pkins

-0.38

startY

-0.38

ImageField

-0.37

postdata

-0.37

ledning

-0.36

 Jeong

-0.36

POSITIVE LOGITS

idespread

1.05

 widely

0.97

 Widely

0.92

 widespread

0.86

wide

0.79

Wide

0.75

 extensively

0.74

广泛

0.68

 ampliamente

0.65

 commonly

0.64

Activations Density 0.013%

expressions indicating widespread use or acceptance

No Comments

No Known Activations