INDEX

Explanations

references to Japan or Japanese culture

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Japanese."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

fy

-0.69

gat

-0.68

 Lyons

-0.67

uliffe

-0.66

 Alexandria

-0.63

allah

-0.63

 Bran

-0.62

 Brees

-0.62

din

-0.62

alla

-0.61

POSITIVE LOGITS

 Japanese

3.69

Japanese

3.35

 Japan

2.55

Japan

2.55

 Taiwanese

2.44

 Chinese

2.21

 Korean

2.21

 Tokyo

2.15

 Indonesian

2.06

 Filipino

2.05

Activations Density 0.012%

references to Japan or Japanese culture

mentions of the word "Japanese."

No Comments

No Known Activations