INDEX

Explanations

country names, especially specific countries in Europe and Asia

oai_token-act-pair · gpt-3.5-turbo

names of countries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Reviewer

-0.89

-0.71

hend

-0.70

é¾įåĸļå£«

-0.68

gru

-0.68

endum

-0.68

senal

-0.66

âĸ¬

-0.66

visor

-0.65

omething

-0.65

POSITIVE LOGITS

ignty

0.81

iphate

0.70

 Taiwan

0.66

 Oman

0.66

uador

0.63

Kel

0.63

 coasts

0.62

 Emirates

0.61

 Guam

0.60

 Brune

0.60

Activations Density 0.235%

country names, especially specific countries in Europe and Asia

names of countries

No Comments

No Known Activations