INDEX

Explanations

references to collective identity or community involvement

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

0:0.03

1:0.02

2:0.18

3:0.09

4:0.11

5:0.06

6:0.04

7:0.07

8:0.08

9:0.06

10:0.12

11:0.09

Negative Logits

 counterparts

-1.64

zu

-1.53

soever

-1.50

umbn

-1.44

 ultras

-1.41

imus

-1.39

wolf

-1.39

sal

-1.38

akedown

-1.37

 Satan

-1.36

POSITIVE LOGITS

.""

1.73

":"/

1.70

EEE

1.58

olor

1.55

ORY

1.52

#$#$

1.49

 spoil

1.47

\.

1.47

LW

1.46

ategory

1.46

Activations Density 0.000%

references to collective identity or community involvement

No Known Activations