INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

è¦ļéĨĴ

-0.76

ãĤ®

-0.73

early

-0.72

chi

-0.70

ãĥīãĥ©

-0.69

ĺ

-0.69

Graph

-0.69

éŃĶ

-0.67

wang

-0.67

ulously

-0.66

POSITIVE LOGITS

 scrimmage

0.82

 forfeiture

0.74

 trench

0.66

 artific

0.65

 trenches

0.65

onies

0.65

 Blair

0.64

imes

0.63

 prest

0.63

 Canter

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.