INDEX

Explanations

references to photographic credits and attributions

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

Photo credits by names

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

engertian

-0.47

klere

-0.45

 læng

-0.42

insegna

-0.42

 puisi

-0.42

 prochaines

-0.41

intenant

-0.41

 bornes

-0.41

 lisäksi

-0.41

 étages

-0.40

POSITIVE LOGITS

Datuak

0.71

OGND

0.68

 المعيارى

0.68

 AssemblyCompany

0.55

 photographs

0.54

 photos

0.54

 photograph

0.52

©

0.52

photograph

0.51

__(/*!

0.51

Activations Density 0.004%

references to photographic credits and attributions

Photo credits by names

No Comments

No Known Activations