INDEX

Explanations

mentions of disabling features or software, particularly related to ad blockers and settings control

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

0:0.02

1:0.01

2:0.06

3:0.09

4:0.15

5:0.03

6:0.04

7:0.29

8:0.03

9:0.04

10:0.11

11:0.08

Negative Logits

代

-1.68

united

-1.59

PsyNetMessage

-1.58

addle

-1.49

ixt

-1.48

quartered

-1.45

 hither

-1.43

owan

-1.42

 circumference

-1.36

worldly

-1.36

POSITIVE LOGITS

 automatic

1.73

NR

1.54

 firewall

1.44

 outright

1.42

 Daylight

1.41

MAP

1.41

 Automatic

1.39

 immunity

1.38

 override

1.36

 False

1.32

Activations Density 0.008%

mentions of disabling features or software, particularly related to ad blockers and settings control

No Known Activations