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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tis

-0.81

LESS

-0.71

âĶĢâĶĢâĶĢâĶĢ

-0.70

Ð²

-0.69

Interstitial

-0.69

Î½

-0.68

DERR

-0.67

Linux

-0.67

BSD

-0.67

âĨĳ

-0.66

POSITIVE LOGITS

uyomi

0.72

umerable

0.68

oys

0.68

odder

0.67

arted

0.66

igious

0.66

 Ducks

0.66

atar

0.65

 Canucks

0.65

aceous

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.