INDEX

Explanations

names

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

words that start with the letter 'J'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 beware

-0.75

 mostly

-0.71

 pony

-0.68

 divided

-0.68

 about

-0.66

 yours

-0.65

 occasionally

-0.65

 Crest

-0.64

 globally

-0.64

 refreshed

-0.64

POSITIVE LOGITS

avascript

1.53

upiter

1.50

asper

1.41

ournals

1.38

ourney

1.38

orge

1.35

avier

1.32

enny

1.30

ihad

1.29

ungle

1.29

Activations Density 0.025%

names

words that start with the letter 'J'

No Comments

No Known Activations