INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Magicka

-0.84

 "$:/

-0.80

 Siren

-0.70

 Sunrise

-0.69

nings

-0.68

 Liquid

-0.67

itialized

-0.66

 Pitch

-0.65

 Devils

-0.65

 League

-0.65

POSITIVE LOGITS

ãĥĺãĥ©

0.81

soDeliveryDate

0.75

century

0.70

%]

0.69

]+

0.64

emp

0.63

arded

0.62

ocide

0.61

 nesting

0.61

]=

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.