INDEX

Explanations

Japanese names and terms, especially related to companies and individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals and companies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

estern

-0.81

arity

-0.75

 Feed

-0.75

tainment

-0.74

ciating

-0.74

aughtered

-0.74

EMENT

-0.73

éĹ

-0.71

ichick

-0.70

OSP

-0.70

POSITIVE LOGITS

ura

0.94

terness

0.91

ury

0.91

ushi

0.90

ub

0.89

uda

0.88

uran

0.86

uki

0.86

uy

0.85

uras

0.82

Activations Density 0.021%

Japanese names and terms, especially related to companies and individuals

proper nouns, particularly names of individuals and companies

No Comments

No Known Activations