INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

PLA

-0.89

Journal

-0.79

Quant

-0.73

Ã©

-0.72

venants

-0.70

Explore

-0.69

break

-0.68

SIM

-0.67

ounds

-0.63

repeat

-0.63

POSITIVE LOGITS

 Thumbnails

0.70

stanbul

0.68

essee

0.67

uther

0.66

 Macedonia

0.66

 Compet

0.65

glers

0.65

 Scand

0.64

bage

0.63

 civilisation

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.