INDEX

Explanations

words related to the term "Wanda."

oai_token-act-pair · gpt-3.5-turbo

words related to specific locations, particularly those ending in "anda"

oai_token-act-pair · gpt-4o-mini

New Auto-Interp

Features by Cosine Similarity

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

Plots

Explanation

Show Test Field

Default Test Text

IFrame

Link

Not in Any Lists

Negative Logits

sed

-1.03

sole

-0.71

alcohol

-0.70

lav

-0.70

bard

-0.69

bred

-0.68

 skelet

-0.68

 deriv

-0.66

dn

-0.66

icka

-0.64

POSITIVE LOGITS

emonium

1.26

 Devi

0.85

 Haram

0.85

ibo

0.83

wana

0.83

ignt

0.79

isance

0.78

uthor

0.76

emon

0.75

arov

0.75

Activations Density 0.011%

words related to the term "Wanda."

words related to specific locations, particularly those ending in "anda"

No Activations Uploaded