INDEX

Explanations

No Explanations Found

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

0:0.07

1:0.07

2:0.07

3:0.07

4:0.08

5:0.09

6:0.07

7:0.07

8:0.09

9:0.09

10:0.08

11:0.08

Negative Logits

 Logo

-3.02

Coy

-2.81

 headlights

-2.77

 Whip

-2.64

 cyclist

-2.56

 Curve

-2.48

 Diesel

-2.46

 windshield

-2.46

 bicycl

-2.43

 pole

-2.42

POSITIVE LOGITS

enza

2.87

entry

2.77

merce

2.68

DragonMagazine

2.63

ibaba

2.54

atche

2.54

Enh

2.48

Pokémon

2.44

enh

2.40

FIELD

2.39

Activations Density 0.000%

No Known Activations

This feature has no known activations.