INDEX

Explanations

the term associated with the football club Borussia Dortmund

oai_token-act-pair · gpt-4-turbo

mentions of the football club "Borussia Dortmund."

oai_token-act-pair · gpt-3.5-turbo

mentions of "Russia."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

changes

-0.81

bor

-0.75

served

-0.72

ably

-0.72

found

-0.72

quickShipAvailable

-0.72

nce

-0.71

tc

-0.70

drawn

-0.70

tailed

-0.70

POSITIVE LOGITS

 Dortmund

1.08

ussia

0.84

 Confederation

0.82

iership

0.78

 Juda

0.78

 Nikola

0.76

sburg

0.75

éĹĺ

0.72

 Region

0.72

 Orchestra

0.71

Activations Density 0.017%

the term associated with the football club Borussia Dortmund

mentions of the football club "Borussia Dortmund."

mentions of "Russia."

No Comments

No Known Activations