INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ï¸ı

-0.75

 Tomorrow

-0.73

 Revival

-0.65

vous

-0.65

 Yesterday

-0.64

 Gorge

-0.64

 Paso

-0.63

 Skies

-0.63

Eag

-0.63

 Hilton

-0.62

POSITIVE LOGITS

ibo

0.90

ritical

0.76

etheless

0.75

tracks

0.74

ert

0.71

 Nanto

0.67

umer

0.66

later

0.65

etr

0.65

ies

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.