INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ô

-0.84

utsche

-0.67

Þ

-0.67

 behavi

-0.65

ibling

-0.65

 offending

-0.65

ļéĨĴ

-0.64

bsite

-0.64

 undermin

-0.63

ñ

-0.62

POSITIVE LOGITS

STON

0.74

facing

0.74

wa

0.72

spir

0.71

radius

0.68

prev

0.67

cum

0.66

uru

0.65

lev

0.65

awks

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.