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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

roup

-0.70

 inflated

-0.66

 Roth

-0.66

 Klein

-0.65

isner

-0.63

 Hein

-0.63

 Zurich

-0.62

paper

-0.61

etting

-0.61

 Burnett

-0.60

POSITIVE LOGITS

pmwiki

0.83

 plaint

0.79

llular

0.71

enos

0.69

tenance

0.68

idav

0.68

Capture

0.67

Compl

0.66

arna

0.66

|--

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.