INDEX

Explanations

directional terms related to left and right

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_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

0:0.31

1:0.05

2:0.05

3:0.04

4:0.07

5:0.05

6:0.14

7:0.04

8:0.03

9:0.02

10:0.11

11:0.03

Negative Logits

at

-1.31

rates

-1.30

atell

-1.29

rad

-1.29

osta

-1.26

rehens

-1.25

aton

-1.25

ato

-1.23

ogi

-1.19

acers

-1.19

POSITIVE LOGITS

 boarding

1.39

 header

1.26

ker

1.22

 traveller

1.21

 Cancel

1.20

footed

1.20

Winged

1.19

 Origin

1.18

 Wander

1.18

 baggage

1.17

Activations Density 0.021%

directional terms related to left and right

No Comments

No Known Activations