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

Recipe

-0.77

-0.76

 Flavoring

-0.73

[&

-0.67

Against

-0.64

erie

-0.63

Reviewer

-0.61

 invited

-0.60

Ingredients

-0.59

 wrench

-0.59

POSITIVE LOGITS

lling

0.79

reditary

0.70

Viz

0.69

Downloadha

0.68

Va

0.67

 Rico

0.67

AAA

0.66

 Cheong

0.65

ugu

0.63

olulu

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.