INDEX

Explanations

various forms of classification and categorization, particularly in relation to groups and organizations

New Auto-Interp

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.02

1:0.02

2:0.28

3:0.18

4:0.06

5:0.05

6:0.05

7:0.03

8:0.05

9:0.09

10:0.07

11:0.05

Negative Logits

 whereas

-1.69

Soc

-1.36

onz

-1.33

Vol

-1.24

because

-1.21

but

-1.20

igr

-1.20

kj

-1.20

bery

-1.15

bian

-1.15

POSITIVE LOGITS

manuel

1.27

�

1.26

aval

1.24

idine

1.23

kees

1.21

estamp

1.21

ゼウス

1.18

\":

1.18

racuse

1.16

icides

1.15

Activations Density 0.012%

various forms of classification and categorization, particularly in relation to groups and organizations

No Known Activations