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

iage

-0.78

 Naples

-0.75

BUG

-0.71

ABE

-0.68

 Towns

-0.65

 Paso

-0.65

 Amer

-0.64

 Nashville

-0.64

BuyableInstoreAndOnline

-0.63

Oro

-0.63

POSITIVE LOGITS

rum

0.70

 contrace

0.67

 consequ

0.67

ordan

0.66

 taught

0.64

anon

0.62

course

0.61

rients

0.61

 thereafter

0.61

oscope

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.