INDEX

Explanations

sentences that conclude a thought or statement

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

1:0.07

2:0.08

3:0.09

4:0.08

5:0.07

6:0.09

7:0.08

8:0.08

9:0.08

10:0.08

11:0.08

Negative Logits

 consultancy

-2.51

 kits

-2.50

 costly

-2.50

kit

-2.49

 publicity

-2.48

 organisation

-2.46

Frameworks

-2.33

 consultation

-2.28

 fundraising

-2.27

 Clar

-2.27

POSITIVE LOGITS

ッド

2.87

ヴ

2.80

nikov

2.77

�

2.74

 Gaul

2.61

igers

2.59

ewitness

2.59

addin

2.55

achus

2.49

ethyst

2.44

Activations Density 0.000%

sentences that conclude a thought or statement

No Known Activations