INDEX

Explanations

words related to catalogs or collections

oai_token-act-pair · gpt-3.5-turbo

mentions of product catalogs or listings

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

Downloadha

-0.95

minent

-0.71

thritis

-0.70

veyard

-0.69

acht

-0.69

adows

-0.65

ulhu

-0.64

wi

-0.64

ths

-0.64

 Outer

-0.62

POSITIVE LOGITS

 catalog

1.28

uing

1.08

 catalogue

1.08

ues

1.04

ued

1.01

 Catalog

0.97

ãĤ¼ãĤ¦ãĤ¹

0.97

ysis

0.97

eers

0.95

ļéĨĴ

0.94

Activations Density 0.005%

words related to catalogs or collections

mentions of product catalogs or listings

No Comments

No Known Activations