INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ichick

-0.83

xon

-0.79

otiation

-0.79

igi

-0.74

ngth

-0.70

enhagen

-0.68

odynam

-0.67

conservancy

-0.66

 aquarium

-0.65

 Advertisement

-0.65

POSITIVE LOGITS

ð

0.67

OHN

0.66

pt

0.64

 guiActiveUnfocused

0.64

arts

0.63

CTV

0.62

ION

0.61

roma

0.61

 cubic

0.60

zon

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.