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

press

-0.72

ureau

-0.61

Pad

-0.60

 Lovecraft

-0.59

ifying

-0.59

 whichever

-0.59

comes

-0.59

DeL

-0.58

IFT

-0.58

Jac

-0.58

POSITIVE LOGITS

 Meal

0.71

ordon

0.70

querque

0.70

ept

0.68

holm

0.68

sted

0.67

eez

0.63

 warr

0.63

xual

0.63

peg

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.