INDEX

Explanations

references to specific historical events or figures

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

 EconPapers

-0.42

ceptos

-0.40

 Burrell

-0.40

 Robin

-0.39

producer

-0.39

liquer

-0.38

puesto

-0.37

 フリース

-0.36

 capacités

-0.36

 frustrated

-0.36

POSITIVE LOGITS

Titanic

0.80

 Titanic

0.76

 AssemblyCulture

0.67

脚注の使い方

0.59

OGND

0.53

titanic

0.52

LookAnd

0.52

itania

0.51

 Picchu

0.51



0.50

Activations Density 0.003%

references to specific historical events or figures

No Comments

No Known Activations