INDEX

Explanations

titles or names that end with 'illi'

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

proper nouns or names associated with specific individuals and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_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.3.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

0:0.02

1:0.03

2:0.05

3:0.34

4:0.03

5:0.02

6:0.05

7:0.19

8:0.07

9:0.04

10:0.07

11:0.05

Negative Logits

BILITIES

-1.65

 Expend

-1.28

ム

-1.26

ATIONS

-1.24

CLASSIFIED

-1.23

��

-1.23

vironment

-1.23

SpaceEngineers

-1.22

VOL

-1.21

デ

-1.20

POSITIVE LOGITS

urat

1.33

angs

1.29

iday

1.23

assetsadobe

1.21

idan

1.19

 wounds

1.19

 wounded

1.15

afa

1.11

ibaba

1.09

beck

1.07

Activations Density 0.027%

titles or names that end with 'illi'

proper nouns or names associated with specific individuals and entities

No Comments

No Known Activations