INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

icz

-0.73

alth

-0.68

Ħ¢

-0.63

ACY

-0.62

 dime

-0.60

stant

-0.58

anus

-0.57

 integers

-0.57

WARD

-0.57

 discouraged

-0.57

POSITIVE LOGITS

audi

0.81

govtrack

0.79

 Ethiop

0.78

PsyNetMessage

0.74

0.70

redit

0.70

natureconservancy

0.67

tag

0.65

RFC

0.64

DOC

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.