INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_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.3.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.08

1:0.09

2:0.08

3:0.09

4:0.07

5:0.09

6:0.08

7:0.07

8:0.08

9:0.06

10:0.08

11:0.09

Negative Logits

arettes

-1.61

.).

-1.43

amera

-1.42

stanbul

-1.40

.):

-1.39

iband

-1.39

chid

-1.39

 syndrome

-1.38

lus

-1.30

′

-1.30

POSITIVE LOGITS

anan

1.57

Chel

1.45

ansky

1.40

POR

1.40

IUM

1.40

Els

1.37

 horizont

1.27

 unden

1.25

Redditor

1.24

 hitherto

1.24

Activations Density 0.000%

No Known Activations

This feature has no known activations.