INDEX

Explanations

the word "Norwegian" in the text

oai_token-act-pair · gpt-3.5-turbo

references to Norway and its cultural or governmental contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

place

-0.80

ually

-0.80

ual

-0.80

plain

-0.80

ept

-0.79

BOOK

-0.74

xxxxxxxx

-0.74

###

-0.72

uers

-0.72

vP

-0.70

POSITIVE LOGITS

Bok

1.07

 Norway

1.05

 Oslo

0.98

 Norwegian

0.97

wegian

0.93

 Refugee

0.84

 Andersen

0.82

borg

0.81

 Stockholm

0.80

Ã¥

0.76

Activations Density 0.008%

the word "Norwegian" in the text

references to Norway and its cultural or governmental contexts

No Comments

No Known Activations