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

âĵĺ

-0.79

 Flavoring

-0.77

Reason

-0.74

Invalid

-0.73

Redd

-0.73

Â§

-0.71

Vert

-0.71

Trivia

-0.68

 Desk

-0.68

Appearances

-0.67

POSITIVE LOGITS

arten

0.77

ornia

0.64

 repr

0.64

 peanuts

0.63

 Clive

0.63

 peer

0.63

 tribute

0.62

udi

0.60

 Kear

0.59

akedown

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.