INDEX

Explanations

formal expressions related to decisions and updates in official contexts

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

0:0.06

1:0.02

2:0.07

3:0.04

4:0.04

5:0.04

6:0.30

7:0.05

8:0.05

9:0.22

10:0.03

11:0.03

Negative Logits

cig

-3.88

 eject

-3.67

 Comcast

-3.65

cery

-3.60

izon

-3.56

 Flavoring

-3.44

Xbox

-3.43

 Barnes

-3.42

izabeth

-3.41

 Grizzlies

-3.38

POSITIVE LOGITS

Ra

9.14

Ra

8.09

ra

5.44

RA

5.30

 Roots

4.26

Ru

4.19

Ri

4.12

 Golem

4.09

ERA

4.04

ra

3.96

Activations Density 0.004%

formal expressions related to decisions and updates in official contexts

No Comments

No Known Activations