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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ry

-0.72

iami

-0.72

 apology

-0.71

 Alexis

-0.67

 Olivier

-0.67

 Wilson

-0.65

 Hernandez

-0.65

 Fraser

-0.64

Pg

-0.63

 Claude

-0.63

POSITIVE LOGITS

ename

0.75

nom

0.73

acts

0.72

Construction

0.70

folk

0.70

Prop

0.68

syn

0.68

phase

0.68

branded

0.67

glomer

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.