INDEX

Explanations

references to "The Wizard of Oz" and related terms like "Wizard," "Oz," or specific character names

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

New Auto-Interp

Top Features by Cosine Similarity

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

0:0.04

1:0.04

2:0.23

3:0.04

4:0.23

5:0.05

6:0.02

7:0.03

8:0.05

9:0.15

10:0.04

11:0.02

Negative Logits

abus

-1.37

 compan

-1.36

actionDate

-1.30

 Kurdish

-1.23

 embargo

-1.23

 fertil

-1.22

 Kurd

-1.22

 foreseeable

-1.22

FSA

-1.22

 volunt

-1.21

POSITIVE LOGITS

 Beasts

1.57

mage

1.51

iuses

1.49

Jr

1.48

merce

1.47

god

1.44

 Wales

1.44

ichick

1.42

 fame

1.40

ドラ

1.40

Activations Density 0.003%

references to "The Wizard of Oz" and related terms like "Wizard," "Oz," or specific character names

No Comments

No Known Activations