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

Redditor

-0.71

Wallet

-0.67

ħĭ

-0.67

ocrates

-0.66

Apps

-0.64

ifle

-0.62

 Annotations

-0.60

 ILCS

-0.59

 Sequence

-0.59

ihara

-0.59

POSITIVE LOGITS

 Vern

0.82

ANCE

0.69

 summit

0.68

insula

0.68

ulent

0.68

ãĥīãĥ©

0.67

ilts

0.65

TBA

0.64

 SOLD

0.63

âĢİ

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.