INDEX

Explanations

references to individuals with roles in an organization or group, specifically assembly members

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

0:0.08

1:0.07

2:0.06

3:0.06

4:0.08

5:0.08

6:0.09

7:0.11

8:0.07

9:0.09

10:0.08

11:0.08

Negative Logits

 Akira

-1.46

 Winc

-1.46

bour

-1.37

 Omega

-1.37

Klu

-1.35

 OPEC

-1.35

BCC

-1.34

 Collider

-1.33

 Corona

-1.33

aurus

-1.32

POSITIVE LOGITS

translation

1.66

 msec

1.48

chieve

1.42

icum

1.40

NECT

1.36

owe

1.34

 sqor

1.33

mem

1.32

morrow

1.31

rar

1.30

Activations Density 0.000%

references to individuals with roles in an organization or group, specifically assembly members

No Known Activations