INDEX

Explanations

references to political entities and actions related to geopolitics

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_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.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.14

1:0.07

2:0.09

3:0.05

4:0.04

5:0.02

6:0.12

7:0.13

8:0.03

9:0.03

10:0.19

11:0.05

Negative Logits

 Monk

-2.77

 seaw

-2.75

Moe

-2.66

 Miracle

-2.65

Magic

-2.65

 Oprah

-2.58

 algae

-2.53

Wa

-2.53

 regener

-2.50

 mercury

-2.48

POSITIVE LOGITS

CIS

9.49

 Cyber

4.18

NI

3.63

CPS

3.61

 Brennan

3.56

DHS

3.35

 Cisco

3.31

CU

3.31

cius

3.28

CS

3.27

Activations Density 0.002%

references to political entities and actions related to geopolitics

No Known Activations