INDEX

Explanations

the word "rather."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

 Williamson

-0.72

 Hermans

-0.69

xff

-0.67

ToBounds

-0.67

 Jackman

-0.66

ctus

-0.65

 Hass

-0.65

 Connelly

-0.64

labelledby

-0.64

 Knowles

-0.64

POSITIVE LOGITS

rather

1.83

 Rather

1.77

Rather

1.77

 rather

1.75

uttosto

1.27

 piuttosto

1.22

 plutôt

1.06

 quite

1.03

Quite

1.00

 fairly

0.99

Activations Density 0.071%

the word "rather."

No Comments

No Known Activations