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

Town

-0.79

ridge

-0.75

Ford

-0.73

gall

-0.72

Miller

-0.71

Whit

-0.70

audi

-0.69

ridges

-0.68

Neg

-0.67

oop

-0.67

POSITIVE LOGITS

 Izan

0.79

 elastic

0.73

 awoken

0.73

 Substance

0.65

 coordination

0.65

 attrition

0.63

 Learns

0.63

 Reincarnated

0.63

esthetic

0.63

 Aviv

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.