INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

reet

-0.74

amous

-0.71

 sidx

-0.66

rider

-0.66

iability

-0.65

crop

-0.65

Dak

-0.64

anda

-0.62

axe

-0.62

oyd

-0.62

POSITIVE LOGITS

puter

0.73

 Solitaire

0.70

 Impro

0.67

 Subtle

0.65

ersen

0.64

 Polic

0.64

 Floating

0.64

 Functional

0.64

aminer

0.63

 Mysterious

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.