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

 Rated

-0.68

 cleaners

-0.63

 refreshed

-0.62

 Irwin

-0.61

 Winner

-0.60

 defective

-0.60

isance

-0.59

 Clever

-0.59

 Frieza

-0.59

 insol

-0.58

POSITIVE LOGITS

oa

0.88

 Cann

0.78

ua

0.73

adelphia

0.72

imore

0.70

ials

0.68

ril

0.68

glers

0.67

eno

0.67

nesota

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.