INDEX

Explanations

references to reading or reading-related activities

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

0:0.07

1:0.08

2:0.07

3:0.09

4:0.08

5:0.08

6:0.09

7:0.09

8:0.09

9:0.07

10:0.07

11:0.08

Negative Logits

 Thursday

-2.72

Vote

-2.66

 Lego

-2.63

ad

-2.61

 Gaga

-2.58

 Wednesday

-2.58

tro

-2.33

 Tuesday

-2.30

 Cruz

-2.30

MC

-2.29

POSITIVE LOGITS

 psychiat

3.37

 inval

3.23

 mosqu

3.15

��

3.03

;;;;;;;;;;;;

3.00

�

2.82

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

2.77

ufact

2.77

 tremend

2.76

 conflic

2.71

Activations Density 0.000%

references to reading or reading-related activities

No Known Activations