INDEX

Explanations

references to rows and datasets in a structured format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 Schweitzer

-0.56

 Bonner

-0.48

 Sutherland

-0.47

-0.46

might

-0.45

 BrowserModule

-0.45

abstract

-0.44

brief

-0.44

 Gottlieb

-0.43

 MIGHT

-0.43

POSITIVE LOGITS

 Rows

1.12

setRow

1.05

 rows

1.01

Row

1.01

Row

0.99

row

0.96

getRow

0.94

 numberOfRows

0.91

Rows

0.89

addRow

0.85

Activations Density 0.171%

references to rows and datasets in a structured format

No Comments

No Known Activations