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

bid

-0.81

eca

-0.72

yout

-0.71

ï¸ı

-0.69

tten

-0.69

clerosis

-0.69

atform

-0.68

gered

-0.67

pter

-0.66

iped

-0.66

POSITIVE LOGITS

 Cable

0.74

Sgt

0.70

SB

0.67

 Princ

0.67

DragonMagazine

0.66

 Bulk

0.64

Compan

0.63

 Xiao

0.63

 Andrews

0.62

 Robbins

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.