INDEX

Explanations

explanations or criticisms related to particular arguments, debates, or evidence

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.07

1:0.08

2:0.06

3:0.08

4:0.03

5:0.08

6:0.09

7:0.06

8:0.06

9:0.12

10:0.16

11:0.06

Negative Logits

 nearby

-1.45

gy

-1.21

 Located

-1.20

Located

-1.16

-1.15

 droid

-1.15

 Moonlight

-1.11

 brakes

-1.09

 Poké

-1.08

 Joker

-1.07

POSITIVE LOGITS

Politics

1.56

 quotations

1.41

soDeliveryDate

1.34

 quotation

1.31

obin

1.31

Polit

1.30

 testimonies

1.30

iberal

1.28

 quoting

1.27

 assertions

1.26

Activations Density 0.621%

explanations or criticisms related to particular arguments, debates, or evidence

No Known Activations