INDEX

Explanations

timestamped social media posts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

posts that emphasize updates and popular interactions or shares

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.06

2:0.08

3:0.07

4:0.03

5:0.12

6:0.03

7:0.04

8:0.06

9:0.30

10:0.06

11:0.03

Negative Logits

 extingu

-1.34

ANG

-1.34

quartered

-1.32

 plutonium

-1.26

 seiz

-1.25

 cooperate

-1.24

gren

-1.23

 Chaser

-1.22

 postwar

-1.19

 warheads

-1.19

POSITIVE LOGITS

 VIDEOS

1.48

nesday

1.39

Bib

1.38

Meta

1.37

 Readers

1.36

 Pastebin

1.36

iquette

1.33

hent

1.28

 Sina

1.27

 reader

1.27

Activations Density 0.074%

timestamped social media posts

posts that emphasize updates and popular interactions or shares

No Comments

No Known Activations