INDEX

Explanations

references to the Quran

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

0:0.08

1:0.09

2:0.07

3:0.07

4:0.08

5:0.07

6:0.08

7:0.08

8:0.09

9:0.08

10:0.08

11:0.08

Negative Logits

jri

-3.21

chuk

-2.97

zh

-2.86

crim

-2.85

agascar

-2.73

anka

-2.71

aiman

-2.59

arijuana

-2.55

ansk

-2.55

 Cambod

-2.54

POSITIVE LOGITS

 Earl

2.71

 inserts

2.41

 SPORTS

2.37

 Polly

2.32

É

2.31

 Doll

2.30

 loads

2.27

 leisure

2.27

 analogue

2.26

Liv

2.25

Activations Density 0.000%

references to the Quran

No Known Activations