INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.9.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

0:0.07

1:0.07

2:0.07

3:0.09

4:0.08

5:0.08

6:0.07

7:0.08

8:0.07

9:0.08

10:0.10

11:0.08

Negative Logits

incial

-2.69

ueless

-2.62

 vetoed

-2.58

ederation

-2.40

oenix

-2.39

inion

-2.37

erald

-2.37

stellar

-2.34

gemony

-2.33

utherford

-2.32

POSITIVE LOGITS

 Nunes

2.73

®

2.65

®,

2.62

 Pamela

2.57

CAM

2.51

 Sakuya

2.43

 Pose

2.42

 Content

2.37

®

2.33

 Paddock

2.33

Activations Density 0.000%

No Known Activations

This feature has no known activations.