INDEX

Explanations

words or abbreviations related to people in politics or influential roles

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

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.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

0:0.03

1:0.03

2:0.03

3:0.07

4:0.02

5:0.04

6:0.16

7:0.04

8:0.04

9:0.40

10:0.03

11:0.04

Negative Logits

Whe

-4.61

Whe

-4.17

Gam

-4.04

whe

-3.80

Wind

-3.55

 Grizzlies

-3.49

Gam

-3.44

Pie

-3.38

whe

-3.37

wig

-3.33

POSITIVE LOGITS

ara

6.87

ARA

5.32

 Sara

5.16

 Kara

4.77

 Tara

4.63

 Lara

4.50

 Mara

4.30

 Clara

4.24

 Paige

3.99

 Barron

3.95

Activations Density 0.002%

words or abbreviations related to people in politics or influential roles

No Comments

No Known Activations