INDEX

Explanations

Japanese names

oai_token-act-pair · gpt-3.5-turbo

names or terms related to Japanese culture or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tering

-0.85

matically

-0.73

sheet

-0.71

laughter

-0.71

ulence

-0.70

trace

-0.70

iaries

-0.70

drivers

-0.69

undy

-0.67

rodu

-0.66

POSITIVE LOGITS

ichi

1.39

 Tsuk

1.34

 Yosh

1.32

oka

1.31

 Tanaka

1.27

ishi

1.26

 Nish

1.24

 Mats

1.24

Tsu

1.21

hiro

1.21

Activations Density 0.136%

Japanese names

names or terms related to Japanese culture or individuals

No Comments

No Known Activations