INDEX

Explanations

phrases indicating a perceived importance or urgency

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_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

0:0.03

1:0.02

2:0.06

3:0.05

4:0.19

5:0.05

6:0.25

7:0.03

8:0.05

9:0.08

10:0.10

11:0.04

Negative Logits

Jed

-1.34

��

-1.31

 Lancaster

-1.30

ogie

-1.22

iste

-1.20

 kicking

-1.17

Pry

-1.15

Mé

-1.14

Kev

-1.14

Ly

-1.13

POSITIVE LOGITS

フォ

1.70

ネ

1.66

ヘラ

1.60

ヴァ

1.58

uble

1.57

ゼウス

1.37

DATA

1.34

ノ

1.32

purpose

1.29

ト

1.28

Activations Density 0.055%

phrases indicating a perceived importance or urgency

No Comments

No Known Activations