INDEX

Explanations

words with the specific suffix "es" in the text

oai_token-act-pair · gpt-3.5-turbo

references to varying concepts or categories

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

DragonMagazine

-0.81

Reviewer

-0.69

CPC

-0.67

tenance

-0.65

 Iranians

-0.64

 Grimes

-0.64

 Russians

-0.64

 Kubrick

-0.63

apple

-0.62

ItemTracker

-0.62

POSITIVE LOGITS

earchers

1.06

oteric

1.06

pec

1.01

terday

0.99

earch

0.96

opher

0.95

sembly

0.89

ierra

0.89

semb

0.87

pect

0.87

Activations Density 0.011%

words with the specific suffix "es" in the text

references to varying concepts or categories

No Comments

No Known Activations