INDEX

Explanations

news-related terms or proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper nouns or significant entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lihood

-0.78

issance

-0.71

åĤ

-0.65

*.

-0.62

uminati

-0.61

insula

-0.61

 Shutterstock

-0.61

ä¹ĭ

-0.60

 curve

-0.60

cise

-0.60

POSITIVE LOGITS

udos

0.73

Net

0.66

 programmers

0.65

lett

0.64

 Brother

0.63

 publishes

0.62

 backers

0.62

Sand

0.61

 wrote

0.61

Cam

0.61

Activations Density 0.462%

news-related terms or proper nouns

proper nouns or significant entities

No Comments

No Known Activations