INDEX

Explanations

historical or controversial information or theories

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.08

1:0.03

2:0.19

3:0.11

4:0.18

5:0.03

6:0.03

7:0.03

8:0.10

9:0.10

10:0.05

11:0.02

Negative Logits

audi

-1.29

 Barron

-1.27

 Frie

-1.26

mun

-1.18

 Indust

-1.16

apest

-1.14

Gou

-1.09

 weeds

-1.08

 Merrill

-1.08

Fre

-1.07

POSITIVE LOGITS

iatus

1.69

aska

1.64

スト

1.53

ciating

1.32

��

1.29

reality

1.29

apesh

1.29

ermanent

1.28

azine

1.26

ipolar

1.22

Activations Density 0.005%

historical or controversial information or theories

No Known Activations