INDEX

Explanations

words and phrases indicating intentionality or deliberation

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

1:0.20

2:0.03

3:0.04

4:0.13

5:0.03

6:0.08

7:0.16

8:0.02

9:0.04

10:0.06

11:0.07

Negative Logits

ailable

-2.84

heid

-2.84

mobile

-2.47

\",

-2.34

\":

-2.34

iven

-2.33

 Gleaming

-2.32

DIV

-2.29

BIL

-2.27

 Stru

-2.21

POSITIVE LOGITS

 unintentional

4.73

 unintentionally

3.85

 unintended

3.60

 intentional

3.43

 accidental

3.36

 inadvert

3.24

 involuntary

2.89

intention

2.88

 intentionally

2.76

 accidentally

2.65

Activations Density 0.001%

words and phrases indicating intentionality or deliberation

No Known Activations