INDEX

Explanations

mentions of various types of shrines

oai_token-act-pair · gpt-3.5-turbo

references to shrines or sacred sites

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

eering

-0.78

 Winchester

-0.74

eers

-0.72

 PowerPoint

-0.70

 laundering

-0.70

 Gutenberg

-0.69

lace

-0.66

 Hebdo

-0.66

POL

-0.65

 Dragonbound

-0.65

POSITIVE LOGITS

anch

1.03

ines

0.95

ower

0.92

shr

0.92

iffs

0.92

inks

0.92

anty

0.90

ith

0.90

ivas

0.89

otted

0.88

Activations Density 0.010%

mentions of various types of shrines

references to shrines or sacred sites

No Comments

No Known Activations