INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 anten

-0.76

 tablets

-0.73

Picture

-0.73

 glim

-0.71

 enthus

-0.70

arians

-0.69

 platoon

-0.68

 tempt

-0.68

 logical

-0.68

 expectancy

-0.67

POSITIVE LOGITS

yre

0.76

leted

0.71

Lago

0.71

Snow

0.66

gur

0.65

cade

0.64

IRO

0.63

athan

0.63

Umb

0.63

 Injury

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.