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

zon

-0.79

 Clicker

-0.76

olson

-0.69

fat

-0.68

itol

-0.68

llor

-0.66

kel

-0.66

roth

-0.66

yang

-0.65

 suspic

-0.65

POSITIVE LOGITS

natureconservancy

0.69

Keys

0.66

itude

0.65

Preview

0.64

Chance

0.63

 arcane

0.63

chemy

0.62

itud

0.62

arter

0.61

ributed

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.