INDEX

Explanations

instances where something is described to be intentionally distorted

oai_token-act-pair · gpt-3.5-turbo

phrases related to opinions on political or environmental issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ModLoader

-0.48

 Seym

-0.43

ody

-0.37

 Pastebin

-0.36

umerable

-0.36

vati

-0.35

awar

-0.35

 carrier

-0.35

agall

-0.35

 vulner

-0.34

POSITIVE LOGITS

 furthermore

0.45

orporated

0.45

 moreover

0.43

bourg

0.42

 UNESCO

0.42

gered

0.42

anwhile

0.41

oths

0.41

enegger

0.40

lycer

0.40

Activations Density 5.201%

instances where something is described to be intentionally distorted

phrases related to opinions on political or environmental issues

No Comments

No Known Activations