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

 Shank

-0.66

 Provision

-0.64

 Matter

-0.64

bridge

-0.63

 place

-0.62

Ky

-0.61

ences

-0.60

Pri

-0.59

Ky

-0.59

 Galile

-0.59

POSITIVE LOGITS

avorite

1.05

rans

0.69

fml

0.69

oled

0.68

ãĤ´ãĥ³

0.68

natureconservancy

0.67

 Niet

0.67

ãĥĥãĥī

0.66

versions

0.66

ãĥĻ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.