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

 Coliseum

-0.74

 removable

-0.70

 Polo

-0.69

 Franch

-0.69

 Karn

-0.68

 Weeks

-0.68

mos

-0.67

Koz

-0.66

Kol

-0.65

 Vengeance

-0.65

POSITIVE LOGITS

IFA

0.81

galitarian

0.74

agonist

0.74

ense

0.69

case

0.68

 ACTIONS

0.67

bery

0.67

omore

0.65

cases

0.65

quest

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.