INDEX

Explanations

negative relationships or associations involving personal dynamics

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

0:0.01

1:0.02

2:0.08

3:0.33

4:0.01

5:0.01

6:0.14

7:0.05

8:0.04

9:0.03

10:0.06

11:0.16

Negative Logits

emis

-1.24

ker

-1.19

inki

-1.15

 hoop

-1.14

ppa

-1.13

eday

-1.10

 Learns

-1.06

 arteries

-1.02

 aggregation

-1.02

ibaba

-1.02

POSITIVE LOGITS

ciating

1.67

�士

1.40

iture

1.37

rences

1.28

�

1.28

foundland

1.26

termination

1.21

�

1.21

opausal

1.20

�

1.19

Activations Density 0.008%

negative relationships or associations involving personal dynamics

No Comments

No Known Activations