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

assic

-0.75

pole

-0.66

phrine

-0.63

winter

-0.61

seless

-0.61

 squads

-0.60

 pole

-0.60

 replacements

-0.59

 replacement

-0.58

 Militia

-0.58

POSITIVE LOGITS

ose

0.73

 curl

0.68

weet

0.65

0.64

 Emin

0.64

ullivan

0.64

 curls

0.64

asio

0.63

å·

0.63

hee

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.