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

 obedient

-0.67

 counted

-0.65

 Shank

-0.62

ALTH

-0.61

 recorded

-0.61

 retired

-0.61

 content

-0.60

 deficit

-0.60

 counting

-0.60

 rested

-0.60

POSITIVE LOGITS

sonian

0.92

Ey

0.77

\":

0.74

liga

0.73

 condem

0.72

enture

0.71

 abdom

0.68

assetsadobe

0.68

sterdam

0.67

forums

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.