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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ripp

-0.72

edia

-0.71

 unsc

-0.64

iph

-0.64

 padd

-0.64

emis

-0.63

 dolphin

-0.62

ret

-0.61

oll

-0.60

amy

-0.60

POSITIVE LOGITS

TextColor

0.83

AppData

0.73

Accessory

0.72

Chance

0.70

 Democr

0.67

btn

0.67

Characters

0.65

 Norn

0.65

ologne

0.65

Dialogue

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.