INDEX

Explanations

instances related to rule violations or disciplinary actions

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

Head Attr Weights

0:0.13

1:0.02

2:0.09

3:0.03

4:0.04

5:0.03

6:0.20

7:0.04

8:0.08

9:0.23

10:0.02

11:0.04

Negative Logits

alg

-4.29

byss

-3.72

cho

-3.70

ugu

-3.68

ENG

-3.62

------------------------------------------------

-3.61

龍契士

-3.48

 Reneg

-3.46

CHO

-3.45

 Gupta

-3.38

POSITIVE LOGITS

 Tate

10.05

 Mare

8.04

 Crate

4.43

 Bates

4.34

 Atkins

4.17

 Barrett

4.10

 Wyatt

4.05

 Tome

4.04

Taj

3.93

 Patterson

3.92

Activations Density 0.002%

instances related to rule violations or disciplinary actions

No Known Activations