INDEX

Explanations

references to violence and police actions during protests and riots

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

1:0.01

2:0.09

3:0.12

4:0.06

5:0.04

6:0.02

7:0.05

8:0.05

9:0.10

10:0.26

11:0.14

Negative Logits

 deserves

-1.32

 Dating

-1.30

thood

-1.29

iverse

-1.26

 Experience

-1.22

inventoryQuantity

-1.19

ormon

-1.18

Experience

-1.17

 destiny

-1.13

 Inher

-1.13

POSITIVE LOGITS

 improvised

1.37

beit

1.32

 gunshots

1.32

ilitary

1.30

 gunfire

1.28

 muzzle

1.27

 retali

1.25

 confisc

1.24

armed

1.24

 passers

1.23

Activations Density 0.543%

references to violence and police actions during protests and riots

No Known Activations