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

 Meridian

-0.75

 Info

-0.73

 Attention

-0.73

PsyNetMessage

-0.72

oids

-0.70

 STATS

-0.70

 reader

-0.67

 Entry

-0.67

 PubMed

-0.66

 Dwell

-0.66

POSITIVE LOGITS

boxing

0.88

sung

0.79

abouts

0.77

pless

0.74

Po

0.68

 administr

0.67

achu

0.66

ssh

0.66

strap

0.66

ende

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.