INDEX

Explanations

phrases indicating ongoing conflicts or controversies

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

0:0.02

1:0.02

2:0.08

3:0.20

4:0.17

5:0.02

6:0.05

7:0.08

8:0.04

9:0.03

10:0.07

11:0.17

Negative Logits

aroo

-1.53

 Polic

-1.50

 Same

-1.49

Laughs

-1.48

 Chains

-1.48

SpaceEngineers

-1.46

hat

-1.44

"/>

-1.44

 Cous

-1.43

Agent

-1.41

POSITIVE LOGITS

 philanthrop

1.43

 mend

1.42

 burgeoning

1.39

 mountainous

1.38

 federally

1.36

 federal

1.36

phabet

1.36

 busy

1.33

 forgiving

1.33

 fertility

1.33

Activations Density 0.013%

phrases indicating ongoing conflicts or controversies

No Comments

No Known Activations