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

 Stras

-0.76

ravel

-0.71

BOX

-0.67

 Coffin

-0.63

 Canaver

-0.63

Ī

-0.63

Ń·

-0.63

 Fernand

-0.62

 compar

-0.62

 Ferdinand

-0.60

POSITIVE LOGITS

lah

0.76

Temp

0.65

âĵĺ

0.65

ixed

0.65

icut

0.65

ratulations

0.64

lations

0.64

otics

0.63

alties

0.63

oud

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.