INDEX

Explanations

phrases that emphasize the importance of respect and dignity in treatment of individuals

New Auto-Interp

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

1:0.02

2:0.07

3:0.06

4:0.08

5:0.01

6:0.04

7:0.51

8:0.02

9:0.02

10:0.05

11:0.06

Negative Logits

whe

-1.68

installed

-1.64

 diagrams

-1.59

stro

-1.51

 footprints

-1.49

rhy

-1.48

uers

-1.48

files

-1.47

ampa

-1.47

andel

-1.46

POSITIVE LOGITS

 lightly

1.66

 respectfully

1.59

 carefully

1.44

PDATE

1.42

Personal

1.42

 fung

1.42

 Admission

1.41

 kindly

1.40

 Honor

1.36

 impart

1.32

Activations Density 0.022%

phrases that emphasize the importance of respect and dignity in treatment of individuals

No Known Activations