INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Liter

-0.70

 subsc

-0.66

weekly

-0.63

Alien

-0.63

incible

-0.63

IRE

-0.62

cigarettes

-0.62

crow

-0.62

ERN

-0.62

á½

-0.62

POSITIVE LOGITS

 Mons

0.75

 consolidation

0.71

 Mirage

0.67

 disg

0.65

NetMessage

0.65

 consolid

0.64

 Franch

0.64

 Marina

0.64

eret

0.63

 memo

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.