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

wcs

-0.72

ORK

-0.69

 Bagg

-0.69

BCC

-0.68

 Cassidy

-0.68

CIS

-0.67

 VIEW

-0.67

 Hughes

-0.67

earch

-0.66

 Codec

-0.65

POSITIVE LOGITS

generation

0.77

dependent

0.73

nom

0.69

borne

0.69

Roman

0.67

based

0.67

gard

0.66

maid

0.65

 pige

0.64

timer

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.