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

 bets

-0.65

 Moons

-0.63

 Rocks

-0.62

intendent

-0.60

 navig

-0.60

ulia

-0.59

 sands

-0.58

 Surprise

-0.58

roleum

-0.57

 Chips

-0.56

POSITIVE LOGITS

 certain

0.82

zin

0.80

TOR

0.78

ENCY

0.72

á¸

0.72

 attRot

0.69

resent

0.68

trak

0.68

icans

0.67

eno

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.