INDEX

Explanations

phrases related to guidance and advice

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

3:0.16

4:0.01

5:0.04

6:0.07

7:0.28

8:0.06

9:0.05

10:0.07

11:0.09

Negative Logits

ikhail

-1.30

etheus

-1.24

allery

-1.18

olina

-1.18

ideo

-1.17

 react

-1.15

ovies

-1.13

 territ

-1.12

 metast

-1.12

hap

-1.11

POSITIVE LOGITS

覚醒

1.29

 Helpful

1.27

 advice

1.21

 Advice

1.19

 recommendation

1.18

 crochet

1.17

 juggling

1.15

 scripture

1.12

 Quran

1.12

 booklet

1.10

Activations Density 0.017%

phrases related to guidance and advice

No Comments

No Known Activations