INDEX

Explanations

religious or moral guidance related words

oai_token-act-pair · gpt-3.5-turbo

verbs and phrases indicative of moral or religious obligations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

tails

-0.79

Wow

-0.75

LS

-0.75

Kids

-0.74

role

-0.72

kids

-0.72

guy

-0.72

EVA

-0.71

ansky

-0.70

licks

-0.70

POSITIVE LOGITS

 esteemed

1.25

 slain

1.23

 reckoned

1.20

 obliged

1.19

 disposed

1.18

 clothed

1.16

 consec

1.14

 acquainted

1.14

 exalted

1.13

 deceived

1.12

Activations Density 0.194%

religious or moral guidance related words

verbs and phrases indicative of moral or religious obligations

No Comments

No Known Activations