INDEX

Explanations

proper nouns related to places or companies

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "New"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

uddin

-0.79

ascript

-0.71

xual

-0.68

ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ

-0.67

ppe

-0.66

ÃĥÃĤ

-0.64

ï¸ı

-0.64

``

-0.63

icone

-0.63

åĮ

-0.63

POSITIVE LOGITS

foundland

1.23

 York

1.17

 Zealand

1.17

bies

1.08

 Orleans

1.02

 YORK

1.01

 Hampshire

0.99

bie

0.98

York

0.95

 Brunswick

0.94

Activations Density 0.046%

proper nouns related to places or companies

occurrences of the word "New"

No Comments

No Known Activations