INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Fed

-0.75

bane

-0.71

anta

-0.71

XM

-0.69

hur

-0.68

rix

-0.68

enda

-0.68

uca

-0.66

ãĤ¸

-0.63

ounty

-0.63

POSITIVE LOGITS

efficients

0.72

elsius

0.71

dit

0.69

tons

0.66

 glim

0.66

 charact

0.64

edes

0.63

 Eden

0.63

 sinks

0.62

 Centauri

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.