INDEX

Explanations

words related to grotesque or horrifying imagery

oai_token-act-pair · gpt-3.5-turbo

terms related to extreme or violent situations

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

BALL

-0.76

horn

-0.72

ACP

-0.72

WARE

-0.68

MQ

-0.66

roads

-0.66

fields

-0.64

 Logged

-0.64

STD

-0.62

wine

-0.62

POSITIVE LOGITS

ities

1.25

ity

1.24

itous

1.18

acies

1.08

inals

1.08

als

1.00

itors

0.98

acy

0.97

ians

0.97

agi

0.96

Activations Density 0.035%

words related to grotesque or horrifying imagery

terms related to extreme or violent situations

No Comments

No Known Activations