INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

erie

-0.81

AIR

-0.73

hell

-0.72

 Sunder

-0.68

 Mehran

-0.67

SHIP

-0.67

forcing

-0.64

zek

-0.64

 Thorn

-0.63

 Sands

-0.62

POSITIVE LOGITS

ffic

0.71

ategor

0.69

 guiActive

0.66

 brightest

0.64

buquerque

0.64

antics

0.63

 Pluto

0.61

icons

0.61

mercial

0.61

ibilities

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.