INDEX

Explanations

entities related to specific geographic locations and institutions

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 avoient

-0.46

 pouvoit

-0.44

 tarvit

-0.40

mişti

-0.40

 vuestros

-0.39

libatkan

-0.38

 auroit

-0.38

 apariencia

-0.38

 rodillas

-0.38

 vuestro

-0.37

POSITIVE LOGITS

 National

0.68

+#+

0.65

 विश्वसनीयता

0.64

 Citizens

0.63

 Joint

0.63

 Jefus

0.62

 Regional

0.62

 kasarigan

0.60

Citizens

0.60

 تضيفلها

0.60

Activations Density 0.445%

entities related to specific geographic locations and institutions

No Comments

No Known Activations