INDEX

Explanations

countries or locations

oai_token-act-pair · gpt-3.5-turbo

mentions of countries

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

erb

-0.70

apple

-0.69

essen

-0.69

upper

-0.68

lihood

-0.67

Sym

-0.67

////

-0.65

ãģŁ

-0.64

Ö¼

-0.63

hooting

-0.63

POSITIVE LOGITS

 invaded

0.86

 Airlines

0.79

's

0.79

 descended

0.76

 annexed

0.70

 governed

0.70

 Arabia

0.69

 legisl

0.69

anasia

0.69

anism

0.68

Activations Density 0.179%

countries or locations

mentions of countries

No Comments

No Known Activations