INDEX

Explanations

descriptions related to artistic or architectural features

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 35-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_35/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.35.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ніципалі

-0.41

новништво

-0.40

 pinulongan

-0.39

 shape

-0.37

脚注の使い方

-0.37

 Esperanto

-0.36

Chham

-0.36

ardust

-0.36

AccessorTable

-0.36

 Writings

-0.35

POSITIVE LOGITS

 landscape

0.62

 landscapes

0.59

Composition

0.58

landscape

0.58

 depicting

0.58

 Landscape

0.58

 portrait

0.58

 Composition

0.58

oil

0.57

Landscape

0.57

Activations Density 0.144%

descriptions related to artistic or architectural features

No Comments

No Known Activations