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

clus

-0.69

 Ware

-0.68

 âĢİ

-0.67

 cruc

-0.67

 Pell

-0.65

âĸ¬

-0.64

 Gate

-0.63

 Sort

-0.62

 Doors

-0.62

 Entry

-0.61

POSITIVE LOGITS

urnal

0.75

pload

0.74

ittle

0.73

imental

0.71

 disapp

0.71

ĸļ

0.68

yip

0.68

ingle

0.67

 acidic

0.67

actionDate

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.