INDEX

Explanations

expressions of opinion and evaluation about various subjects

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

Head Attr Weights

0:0.01

1:0.01

2:0.24

3:0.10

4:0.17

5:0.01

6:0.05

7:0.10

8:0.03

9:0.03

10:0.10

11:0.08

Negative Logits

athlon

-1.28

 registry

-1.26

alions

-1.26

Hue

-1.22

 contingency

-1.20

ESS

-1.18

Narr

-1.18

 hormonal

-1.17

 hazardous

-1.14

ormonal

-1.14

POSITIVE LOGITS

 congr

1.82

especially

1.41

{:

1.39

 Pengu

1.33

 honesty

1.33

 considering

1.32

eln

1.28

opian

1.25

ertodd

1.24

thank

1.23

Activations Density 0.283%

expressions of opinion and evaluation about various subjects

No Known Activations