INDEX

Explanations

various forms of the word "ad" or references to advertisements

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.02

2:0.05

3:0.06

4:0.06

5:0.05

6:0.41

7:0.05

8:0.05

9:0.06

10:0.07

11:0.04

Negative Logits

leep

-1.44

¯

-1.42

usercontent

-1.41

icked

-1.31

 Hallow

-1.23

 seniors

-1.20

 compuls

-1.19

sed

-1.16

etheless

-1.15

 Jolly

-1.14

POSITIVE LOGITS

igham

1.74

utical

1.45

IRO

1.44

ciation

1.36

iscons

1.35

ignt

1.33

emale

1.30

ダ

1.30

rouse

1.28

combe

1.27

Activations Density 0.002%

various forms of the word "ad" or references to advertisements

No Comments

No Known Activations