INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 guiActiveUn

-0.86

ibaba

-0.83

 seiz

-0.77

pez

-0.71

ials

-0.70

estate

-0.67

 ribbon

-0.66

kit

-0.66

 spacing

-0.66

————————

-0.64

POSITIVE LOGITS

ULL

0.70

POSE

0.68

 Okawaru

0.67

 Toad

0.65

ãĥĪ

0.65

Graph

0.64

Kill

0.64

 Rodney

0.64

 Tinker

0.62

Shy

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.