INDEX

Explanations

references to the city of Kraków

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_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_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.10.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

0:0.27

1:0.21

2:0.03

3:0.02

4:0.03

5:0.02

6:0.05

7:0.07

8:0.03

9:0.04

10:0.15

11:0.04

Negative Logits

 Watson

-4.00

 Mellon

-3.58

 Chong

-3.24

 Camb

-3.17

 Howell

-3.16

 Dian

-3.15

 Hubbard

-3.15

 Scient

-3.08

 Nero

-3.08

 Corvette

-3.00

POSITIVE LOGITS

rak

6.90

Rah

4.05

 Perse

3.81

arak

3.73

 Ariel

3.40

Rag

3.39

 paras

3.34

 Rampage

3.33

rake

3.31

Rez

3.30

Activations Density 0.001%

references to the city of Kraków

No Comments

No Known Activations