INDEX

Explanations

news-related words and phrases

oai_token-act-pair · gpt-3.5-turbo

instances of end-of-text markers, indicating the conclusion of segments or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 manif

-0.54

sic

-0.53

iqueness

-0.47

bom

-0.47

eday

-0.46

 nearest

-0.45

Shape

-0.45

akespe

-0.45

aturdays

-0.44

ãĥ¼ãĥĨãĤ£

-0.43

POSITIVE LOGITS

Streamer

0.58

 âĢº

0.57

 largeDownload

0.54

PRESS

0.51

 ARTICLE

0.49

 ILCS

0.49

Associated

0.47

 GOODMAN

0.47

UTERS

0.46

PORT

0.46

Activations Density 0.564%

news-related words and phrases

instances of end-of-text markers, indicating the conclusion of segments or statements

No Comments

No Known Activations