INDEX

Explanations

numerical information presented in the format '[number]'

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, specifically periods and brackets

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

 nearest

-0.67

ãĥ¼ãĥĨ

-0.65

 intensive

-0.65

 intellig

-0.63

 nearby

-0.63

 indust

-0.63

 collide

-0.62

arettes

-0.62

ãĥ¼ãĥ³

-0.62

 predec

-0.62

POSITIVE LOGITS

...]

1.51

âĢ¦]

1.43

Pg

1.34

?]

1.22

emphasis

1.06

!]

1.03

note

1.02

sic

1.00

src

0.99

].

0.97

Activations Density 0.038%

numerical information presented in the format '[number]'

punctuation marks, specifically periods and brackets

No Comments

No Known Activations