INDEX

Explanations

No Explanations Found

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sanity

-0.68

oming

-0.65

 arsen

-0.64

 respir

-0.63

 safest

-0.62

orable

-0.61

LED

-0.61

opot

-0.60

 caut

-0.60

eas

-0.60

POSITIVE LOGITS

trump

0.75

å§«

0.67

respond

0.64

),"

0.63

dos

0.62

pipe

0.61

 Whitman

0.61

Cab

0.61

yip

0.60

 Totem

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.