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

prototype

-0.78

roads

-0.77

odore

-0.73

amar

-0.72

aster

-0.71

utenberg

-0.69

 laun

-0.66

 quadru

-0.64

rary

-0.62

-0.61

POSITIVE LOGITS

 Shades

0.73

iciary

0.71

ghai

0.69

ction

0.69

KK

0.67

 McCann

0.66

 Rivers

0.65

Kings

0.65

ãĤ©

0.61

 groom

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.