INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Enlarge

-0.76

 Polish

-0.69

HDR

-0.68

ÅĤ

-0.68

CPR

-0.67

rador

-0.66

NTS

-0.65

 Spicer

-0.64

PsyNetMessage

-0.62

 retina

-0.61

POSITIVE LOGITS

oulos

0.75

andise

0.75

munition

0.69

luster

0.67

bang

0.67

 cumbers

0.67

 unintention

0.66

aditional

0.64

cdn

0.64

 inherent

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.