INDEX

Explanations

expressions of disbelief or surprise regarding societal issues

New Auto-Interp

Top Features by Cosine Similarity

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

0:0.01

1:0.01

2:0.10

3:0.11

4:0.23

5:0.01

6:0.04

7:0.22

8:0.03

9:0.03

10:0.04

11:0.12

Negative Logits

thood

-1.69

imize

-1.59

apon

-1.56

oneliness

-1.55

avorite

-1.54

roxy

-1.49

icter

-1.49

-1.47

hetically

-1.45

leep

-1.42

POSITIVE LOGITS

 Anon

1.48

 Laun

1.42

Led

1.38

Mus

1.35

Pon

1.33

 Berk

1.32

Ely

1.32

 Faculty

1.31

Led

1.31

Ans

1.23

Activations Density 0.004%

expressions of disbelief or surprise regarding societal issues

No Known Activations