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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Tavern

-0.70

 Odyssey

-0.70

 Gent

-0.70

 Juda

-0.70

Io

-0.70

 Hond

-0.69

 Masquerade

-0.69

DC

-0.69

Orc

-0.68

Tue

-0.68

POSITIVE LOGITS

 therap

0.81

govtrack

0.76

userc

0.75

igslist

0.75

monary

0.73

ichita

0.72

 pestic

0.72

ancy

0.72

ettings

0.71

glomer

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.