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

Vu

-0.81

 Siem

-0.73

atron

-0.73

PNG

-0.70

 Watts

-0.69

 Manson

-0.67

obiles

-0.65

 Euros

-0.64

 Immunity

-0.64

 Generation

-0.64

POSITIVE LOGITS

emis

0.88

bid

0.85

pherd

0.79

cyclopedia

0.78

xual

0.78

hetical

0.78

Adds

0.77

eus

0.77

Sensor

0.76

iaries

0.76

Activations Density 0.000%

No Known Activations

This feature has no known activations.