INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

DERR

-0.76

Plan

-0.69

 Modes

-0.67

 Tail

-0.66

 Invasion

-0.65

Mania

-0.64

 Pixie

-0.64

 Brands

-0.62

 Wave

-0.61

Mach

-0.61

POSITIVE LOGITS

orne

0.73

 captcha

0.70

utsche

0.69

 indemn

0.68

 mystery

0.67

 largeDownload

0.65

icted

0.64

pse

0.64

vou

0.64

nig

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.