INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Typhoon

-0.76

 Fighter

-0.72

pload

-0.67

 Pastebin

-0.64

 calib

-0.64

fam

-0.64

 RELE

-0.63

medi

-0.63

xf

-0.63

ichi

-0.63

POSITIVE LOGITS

pherd

0.74

Mars

0.73

eus

0.72

uggest

0.69

istically

0.69

que

0.68

iseum

0.66

 Mars

0.65

zsche

0.64

essee

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.