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

default

-0.67

Application

-0.66

llah

-0.65

 Attribution

-0.63

ses

-0.62

 additionally

-0.61

kus

-0.60

COM

-0.60

utive

-0.60

gas

-0.60

POSITIVE LOGITS

terday

0.85

quickShipAvailable

0.81

DERR

0.72

ammers

0.65

orb

0.64

enery

0.63

Osw

0.62

 defenders

0.61

 banners

0.61

ipers

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.