INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Feldman

-0.74

haar

-0.70

 detached

-0.69

 trump

-0.65

 sway

-0.65

 voic

-0.64

 fused

-0.64

 disg

-0.63

charg

-0.62

«

-0.61

POSITIVE LOGITS

rast

0.83

rol

0.83

bernatorial

0.73

advertisement

0.72

aunted

0.71

uristic

0.71

rolled

0.70

icipated

0.70

Reviewer

0.68

sov

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.