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

opol

-0.68

dis

-0.67

ãĤ¦

-0.67

vre

-0.64

×¨

-0.63

ographical

-0.62

perties

-0.62

ol

-0.60

uebl

-0.60

ruct

-0.60

POSITIVE LOGITS

sung

0.76

abase

0.74

 trave

0.68

mber

0.67

Peg

0.67

 Sweeney

0.66

eva

0.66

antine

0.65

 leash

0.64

Reward

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.