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

ipel

-0.70

resses

-0.67

Jew

-0.64

arget

-0.62

Roz

-0.62

pixel

-0.60

Cover

-0.60

åĨ

-0.58

ceans

-0.58

faces

-0.58

POSITIVE LOGITS

erous

0.69

...]

0.67

 scoreboard

0.66

 trave

0.65

cial

0.64

 backdoor

0.63

 Roose

0.63

 cooker

0.62

court

0.62

quickShipAvailable

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.