INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of specific entities or concepts, particularly proper nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

cop

-0.79

 Broad

-0.78

ipl

-0.76

Wiki

-0.76

 Plate

-0.74

Bi

-0.72

Bio

-0.71

bi

-0.69

Cy

-0.69

 Quad

-0.68

POSITIVE LOGITS

ãĤĭ

2.99

ãģ¦

2.78

ãģŁ

2.70

ãģ¾

1.84

ãģª

1.67

ãĤī

1.52

ãĤĮ

1.37

 contr

1.30

ãģı

1.28

 Johnston

1.21

Activations Density 0.043%

proper nouns

repeated mentions of specific entities or concepts, particularly proper nouns

No Comments

No Known Activations