INDEX

Explanations

mentions of specific countries, with a focus on Ethiopia

oai_token-act-pair · gpt-3.5-turbo

references to Ethiopia

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rals

-0.83

negie

-0.78

aries

-0.76

angular

-0.72

iard

-0.72

ership

-0.70

arten

-0.69

uscript

-0.69

clips

-0.67

uli

-0.67

POSITIVE LOGITS

 Ethiopian

0.89

 Ethiopia

0.74

 Ethiop

0.74

Ò

0.74

udeb

0.68

ener

0.63

 Typhoon

0.63

Micro

0.63

 Giov

0.62

Kut

0.61

Activations Density 0.029%

mentions of specific countries, with a focus on Ethiopia

references to Ethiopia

No Comments

No Known Activations