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

SHIP

-0.71

xp

-0.69

 Skies

-0.68

xon

-0.67

sels

-0.65

SEA

-0.64

 drones

-0.63

deals

-0.61

DHS

-0.58

bike

-0.58

POSITIVE LOGITS

ibrary

0.71

itton

0.65

ritic

0.65

abled

0.62

Com

0.62

agara

0.62

 cryst

0.59

urdy

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.