INDEX

Explanations

phrases related to legal and authoritative matters

oai_token-act-pair · gpt-3.5-turbo

references to authorship and trademarks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Copyright

-0.56

horn

-0.53

 guiActiveUn

-0.51

ãĤµ

-0.51

des

-0.51

Flickr

-0.50

azeera

-0.49

enn

-0.48

ÂŃ

-0.48

cult

-0.47

POSITIVE LOGITS

 withd

0.68

anwhile

0.64

 challeng

0.59

rul

0.58

 horizont

0.58

æĸ¹

0.56

alion

0.56

 Antar

0.56

 mosqu

0.55

ATHER

0.55

Activations Density 2.051%

phrases related to legal and authoritative matters

references to authorship and trademarks

No Comments

No Known Activations