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

jriwal

-0.74

RFC

-0.70

 elevation

-0.69

Cod

-0.69

Sym

-0.68

itol

-0.67

Color

-0.65

*/(

-0.64

Override

-0.64

votes

-0.64

POSITIVE LOGITS

 Mercy

0.75

iors

0.69

urch

0.68

isbury

0.63

 voic

0.62

HUD

0.61

Aid

0.61

 drawer

0.61

isons

0.60

 Dres

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.