INDEX

Explanations

instances of the word "ten" in the text

oai_token-act-pair · gpt-3.5-turbo

references to the number ten

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

hammad

-0.69

 DRAG

-0.69

Downloadha

-0.63

 jerk

-0.63

à¼

-0.61

 Indust

-0.60

 Bind

-0.59

 Tonight

-0.59

 depri

-0.57

HY

-0.57

POSITIVE LOGITS

aciously

1.31

acious

1.21

acity

1.17

ured

1.10

uous

1.09

uously

1.02

 thousand

1.00

fold

1.00

ements

0.99

eenth

0.96

Activations Density 0.010%

instances of the word "ten" in the text

references to the number ten

No Comments

No Known Activations