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

awk

-0.72

alid

-0.70

cerned

-0.69

tera

-0.65

Moj

-0.65

anas

-0.65

Target

-0.64

LU

-0.63

 Payton

-0.63

olars

-0.63

POSITIVE LOGITS

ikuman

0.72

NH

0.63

 âĨĳ

0.63

 altogether

0.62

worthy

0.62

wax

0.61

Gum

0.61

ored

0.60

 Antarctica

0.60

âĶĢâĶĢ

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.