INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

UGC

-0.80

aughs

-0.76

æŃ¦

-0.70

ridor

-0.70

PHOTOS

-0.69

çĶ°

-0.69

ä½ľ

-0.65

 Fargo

-0.64

 Pistons

-0.63

raltar

-0.63

POSITIVE LOGITS

tein

0.80

 stool

0.69

edom

0.69

nutrition

0.69

illion

0.67

ule

0.67

pop

0.67

soever

0.67

ected

0.66

 invaders

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.