INDEX

Explanations

clothing items, specifically T-shirts

oai_token-act-pair · gpt-3.5-turbo

references to T-shirts

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

Reviewer

-0.77

audi

-0.71

«ĺ

-0.70

ntil

-0.69

judicial

-0.66

 presided

-0.63

Aud

-0.62

ļ

-0.62

 distant

-0.62

moving

-0.61

POSITIVE LOGITS

shirt

1.42

shirts

1.28

hirt

1.09

 shirts

1.06

 shirt

0.97

leeve

0.89

 Shirt

0.88

idas

0.88

cloth

0.86

 sleeves

0.86

Activations Density 0.008%

clothing items, specifically T-shirts

references to T-shirts

No Comments

No Known Activations