INDEX

Explanations

names and terms with varying levels of prominence or significance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

Not in Any Lists

No Comments

Head Attr Weights

0:0.04

1:0.03

2:0.03

3:0.09

4:0.02

5:0.04

6:0.15

7:0.03

8:0.05

9:0.44

10:0.02

11:0.02

Negative Logits

 Webs

-3.81

Tul

-3.74

 glyph

-3.63

tul

-3.50

Kem

-3.33

 Troy

-3.31

 telesc

-3.31

 webs

-3.31

 Graves

-3.28

 Strongh

-3.19

POSITIVE LOGITS

ino

6.78

INO

6.03

inos

5.11

ano

4.95

ovo

4.62

ono

4.27

appa

4.22

 Dino

4.12

iano

4.12

apon

3.99

Activations Density 0.002%

names and terms with varying levels of prominence or significance

No Comments

No Known Activations