INDEX

Explanations

phrases related to future plans or predictions

oai_token-act-pair · gpt-3.5-turbo

expressions related to influence and decision-making

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

 indo

-0.34

)."

-0.32

 Afgh

-0.31

"/>

-0.31

]."

-0.30

 disadvant

-0.30

 vulner

-0.30

 unemploy

-0.29

 destro

-0.29

 undermin

-0.29

POSITIVE LOGITS

ivating

0.33

utterstock

0.32

ideshow

0.31

agonist

0.31

urable

0.30

asting

0.30

eaturing

0.30

heimer

0.29

iven

0.29

ering

0.29

Activations Density 3.802%

phrases related to future plans or predictions

expressions related to influence and decision-making

No Comments

No Known Activations