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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 unden

-0.81

 Canaver

-0.73

 Kund

-0.71

icter

-0.68

Sov

-0.66

Eck

-0.65

 purified

-0.65

 unification

-0.65

 Farn

-0.63

ngth

-0.63

POSITIVE LOGITS

mercial

0.75

uca

0.73

ascript

0.70

wal

0.69

teenth

0.67

merce

0.67

lectic

0.65

mouth

0.65

alos

0.64

aucus

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.