INDEX

Explanations

geographic locations, particularly cities, states, and countries

oai_token-act-pair · gpt-3.5-turbo

distinctive names or terms often associated with specific identities or locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

pick

-0.80

picking

-0.78

tr

-0.77

space

-0.75

lot

-0.74

short

-0.71

mind

-0.70

scale

-0.69

tons

-0.67

POSITIVE LOGITS

ÅŁ

1.41

ÄŁ

1.27

Äĩ

1.26

uthor

1.26

eva

1.23

Ã±

1.23

emon

1.21

udeau

1.20

Ã¯

1.19

usk

1.17

Activations Density 0.231%

geographic locations, particularly cities, states, and countries

distinctive names or terms often associated with specific identities or locations

No Comments

No Known Activations