INDEX

Explanations

phrases related to assigning blame or responsibility

oai_token-act-pair · gpt-3.5-turbo

phrases indicating classification or attribution of characteristics and roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

meric

-0.77

PsyNetMessage

-0.73

strap

-0.70

oop

-0.70

brance

-0.69

ovember

-0.68

aceous

-0.68

 stripe

-0.64

irin

-0.62

sche

-0.62

POSITIVE LOGITS

«ĺ

0.70

qqa

0.65

 modem

0.62

 Watergate

0.60

 depl

0.54

wil

0.54

 Cunningham

0.54

Ùħ

0.53

Ĥ¬

0.52

 Batt

0.52

Activations Density 0.746%

phrases related to assigning blame or responsibility

phrases indicating classification or attribution of characteristics and roles

No Comments

No Known Activations