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

shape

-0.72

perm

-0.70

uyomi

-0.70

 amber

-0.68

 specialize

-0.67

 strap

-0.67

 touched

-0.67

ensical

-0.64

 Anon

-0.64

 untouched

-0.64

POSITIVE LOGITS

ang

1.75

angs

1.08

ãĤ¼

0.86

ANG

0.83

 Scully

0.79

angan

0.77

yu

0.72

ei

0.72

grim

0.71

Ba

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.