INDEX

Explanations

references to books, awards, and related scientific terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

ate

-0.71

ent

-0.69

ale

-0.69

ant

-0.68

ar

-0.68

ati

-0.68

ak

-0.68

el

-0.68

ris

-0.67

ang

-0.66

POSITIVE LOGITS

 Houſe

0.61

 Majefty

0.57

 wikipagina

0.55

 AssemblyCulture

0.54

 houſe

0.54

 faſt

0.54

 ainfi

0.53

 vectoriales

0.53

ambién

0.53

 Wikiseite

0.51

Activations Density 3.698%

references to books, awards, and related scientific terms

No Comments

No Known Activations