INDEX

Explanations

references to discrimination and social issues

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.07

1:0.04

2:0.04

3:0.05

4:0.09

5:0.03

6:0.08

7:0.21

8:0.02

9:0.04

10:0.04

11:0.24

Negative Logits

 mystery

-2.36

 illusion

-2.33

illusion

-2.32

 myth

-2.27

Shot

-2.17

isma

-2.14

 scheme

-2.14

rast

-2.12

 reward

-2.11

 courier

-2.10

POSITIVE LOGITS

 remarks

3.00

 comments

2.72

Comments

2.50

 statements

2.46

 homophobic

2.39

 observations

2.33

 derogatory

2.30

izations

2.29

comments

2.26

WC

2.20

Activations Density 0.015%

references to discrimination and social issues

No Known Activations