INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

chance

-0.72

²¾

-0.70

inately

-0.67

etheus

-0.67

 count

-0.65

ulence

-0.65

ulent

-0.65

inel

-0.63

paren

-0.63

 case

-0.62

POSITIVE LOGITS

 Canaver

0.79

 Carney

0.68

 Laws

0.66

 Bris

0.65

Faw

0.65

 Diet

0.64

 Christy

0.64

 Soda

0.63

 Galile

0.63

poons

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.