INDEX

Explanations

mentions of the term "Japanese"

oai_token-act-pair · gpt-3.5-turbo

repeated references to "Japanese."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

lain

-0.90

onies

-0.88

mble

-0.88

izons

-0.85

rost

-0.83

rent

-0.79

staking

-0.78

ithmetic

-0.77

Ì¶

-0.76

chel

-0.76

POSITIVE LOGITS

yen

1.09

Yen

1.01

Japanese

0.81

imura

0.79

istani

0.79

oka

0.77

 Earthquake

0.75

 Â¥

0.75

£ı

0.74

 Japanese

0.73

Activations Density 0.009%

mentions of the term "Japanese"

repeated references to "Japanese."

No Comments

No Known Activations