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

 seeker

-0.67

MPH

-0.67

 screenings

-0.67

 braking

-0.65

 screening

-0.63

 fluorescent

-0.62

icum

-0.62

ierrez

-0.60

çİĭ

-0.59

UID

-0.59

POSITIVE LOGITS

iatus

0.78

 Nost

0.72

anto

0.69

abeth

0.68

issance

0.67

 Estate

0.65

 Vance

0.65

 Aunt

0.63

rox

0.63

bj

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.