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

wolf

-0.68

ometry

-0.67

 dstg

-0.67

GD

-0.64

 traff

-0.64

 Jinn

-0.64

 uncomp

-0.62

 strugg

-0.62

 psychiat

-0.62

 taxp

-0.61

POSITIVE LOGITS

âĵĺ

0.78

ãĤ´

0.72

ocalypse

0.70

rance

0.68

uku

0.66

rue

0.65

arta

0.63

ordon

0.61

myra

0.60

tesy

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.