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

 Luthor

-0.71

 Ribbon

-0.67

ruciating

-0.67

 targ

-0.64

CoC

-0.62

 veter

-0.61

 Bind

-0.61

 superpower

-0.61

Diamond

-0.61

CLASSIFIED

-0.61

POSITIVE LOGITS

ews

0.82

SourceFile

0.74

OPA

0.73

awks

0.72

Ô

0.71

anu

0.67

nings

0.65

orts

0.65

anders

0.64

RT

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.