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

 corrid

-0.77

iHUD

-0.76

 Sutherland

-0.70

 improv

-0.68

ãĤº

-0.66

 Andersen

-0.64

assic

-0.64

CCP

-0.63

¬¼

-0.63

 audiences

-0.62

POSITIVE LOGITS

watch

0.77

chlor

0.77

gee

0.73

cgi

0.70

need

0.68

jay

0.67

 Flow

0.66

bound

0.66

operation

0.65

conduct

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.