INDEX

Explanations

rankings or positions in competitions

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_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

0:0.02

1:0.02

2:0.06

3:0.06

4:0.29

5:0.02

6:0.07

7:0.13

8:0.02

9:0.02

10:0.11

11:0.13

Negative Logits

 likeness

-1.26

 metaphor

-1.25

ument

-1.24

 applic

-1.21

 hindsight

-1.18

 avail

-1.18

 alleg

-1.17

 pause

-1.16

 subscription

-1.15

 appeal

-1.15

POSITIVE LOGITS

HW

1.45

ecause

1.44

busters

1.30

 Overall

1.28

 Commands

1.26

 tallest

1.24

 Breed

1.24

�

1.23

cats

1.23

none

1.20

Activations Density 0.018%

rankings or positions in competitions

No Comments

No Known Activations