INDEX

Explanations

mentions of the word "Japanese"

oai_token-act-pair · gpt-3.5-turbo

references to Japan or Japanese entities

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

lier

-0.89

riers

-0.86

mble

-0.86

vable

-0.82

rent

-0.81

byss

-0.80

staking

-0.80

ttes

-0.79

lain

-0.78

izons

-0.76

POSITIVE LOGITS

Yen

1.04

yen

0.95

 embassy

0.81

 Instruments

0.78

 nationals

0.78

oka

0.77

apolis

0.75

 consulate

0.74

ya

0.74

£ı

0.73

Activations Density 0.008%

mentions of the word "Japanese"

references to Japan or Japanese entities

No Comments

No Known Activations