INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

alysed

-0.80

zos

-0.76

glas

-0.72

hest

-0.68

 sway

-0.66

ãĤ»

-0.66

azar

-0.65

 elect

-0.64

vati

-0.64

chens

-0.63

POSITIVE LOGITS

Vill

0.68

ydia

0.66

 Rookie

0.60

Args

0.59

EFF

0.59

 Ryder

0.58

OPA

0.57

MFT

0.57

 Heroic

0.57

OTH

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.