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

é¾įå¥ĳå£«

-0.79

pter

-0.66

 Teach

-0.66

gear

-0.65

 Showdown

-0.63

advant

-0.61

antz

-0.61

 opponent

-0.60

-0.59

romeda

-0.59

POSITIVE LOGITS

ãĥĢ

0.75

leys

0.67

vation

0.64

cule

0.63

atel

0.62

unal

0.62

Row

0.62

ibel

0.61

cious

0.61

ams

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.