INDEX

Explanations

phrases related to essential items or products commonly used or consumed

oai_token-act-pair · gpt-3.5-turbo

references to essential or common items and concepts, often labeled as "staples."

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

 Galile

-0.81

Zero

-0.75

Zer

-0.72

Aj

-0.72

zero

-0.70

akia

-0.68

Cross

-0.66

 Pirate

-0.66

eton

-0.65

VPN

-0.64

POSITIVE LOGITS

 Staples

4.00

aples

2.43

 staples

2.07

 staple

1.67

 Eaton

1.22

 Lowe

1.19

 Sears

1.16

 Crate

1.00

 Goods

0.99

dur

0.99

Activations Density 0.051%

phrases related to essential items or products commonly used or consumed

references to essential or common items and concepts, often labeled as "staples."

No Comments

No Known Activations