INDEX

Explanations

references to deep philosophical concepts or profound truths

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.06

1:0.06

2:0.09

3:0.05

4:0.03

5:0.10

6:0.03

7:0.07

8:0.06

9:0.27

10:0.08

11:0.05

Negative Logits

anwhile

-1.40

aution

-1.37

 deployment

-1.26

 deployments

-1.26

 principals

-1.20

 Implementation

-1.19

 behavi

-1.17

 escalate

-1.16

 looph

-1.16

QB

-1.15

POSITIVE LOGITS

 pilgr

1.36

 Lovecraft

1.35

iru

1.34

 mystic

1.32

 Grail

1.30

 Sorce

1.30

 Metatron

1.29

 Zoro

1.29

 Sora

1.28

 Souls

1.27

Activations Density 0.006%

references to deep philosophical concepts or profound truths

No Known Activations