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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

éļ

-0.74

ortmund

-0.71

acan

-0.68

 closures

-0.67

tip

-0.67

 tipping

-0.66

te

-0.66

skirts

-0.64

 referees

-0.64

osher

-0.63

POSITIVE LOGITS

(-

0.75

rious

0.72

yll

0.67

naire

0.66

 Sylv

0.65

ÂłÂłÂłÂł

0.64

UTH

0.62

 Louie

0.62

EngineDebug

0.62

 Labrador

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.