INDEX

Explanations

phrases indicating various types of substances or classifications

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

2:0.13

3:0.07

4:0.06

5:0.05

6:0.11

7:0.22

8:0.04

9:0.04

10:0.08

11:0.10

Negative Logits

oppable

-1.44

anwhile

-1.38

 Guth

-1.37

ournal

-1.36

 Johannes

-1.33

ancies

-1.32

uthor

-1.29

vertisements

-1.28

 laun

-1.28

 occas

-1.27

POSITIVE LOGITS

RNA

1.52

RN

1.38

 Common

1.34

audi

1.30

Http

1.29

examination

1.28

 antit

1.25

DragonMagazine

1.22

VC

1.21

resent

1.20

Activations Density 0.004%

phrases indicating various types of substances or classifications

No Comments

No Known Activations