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

laus

-0.69

PLIED

-0.68

SPONSORED

-0.68

INT

-0.63

inas

-0.63

usr

-0.63

 Siren

-0.63

["

-0.62

âĨĳ

-0.61

figure

-0.60

POSITIVE LOGITS

ãĤ¤ãĥĪ

0.73

 Dracula

0.72

 Awards

0.69

aration

0.63

 Opportun

0.63

 cater

0.62

 Action

0.61

 brist

0.60

deals

0.60

agonist

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.