INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

hops

-0.71

 coordination

-0.64

 facult

-0.64

 tempor

-0.62

arcity

-0.62

 spont

-0.62

 philosophers

-0.59

lopp

-0.59

ioned

-0.58

ierre

-0.57

POSITIVE LOGITS

Cu

0.74

athan

0.69

 Nero

0.67

CHA

0.66

erno

0.65

TN

0.65

OVA

0.64

Beat

0.64

shock

0.63

ļéĨĴ

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.