INDEX

Explanations

repeated use of the word "Polis" and variations of "ikarpov" alongside other related terms

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_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

0:0.02

1:0.03

2:0.12

3:0.21

4:0.02

5:0.02

6:0.14

7:0.09

8:0.04

9:0.08

10:0.07

11:0.10

Negative Logits

 Hunger

-1.13

gerald

-1.10

 Franch

-1.07

 collisions

-1.03

 Coliseum

-1.02

igible

-1.02

 Burke

-0.98

 Penalty

-0.97

 MacDonald

-0.96

ourage

-0.95

POSITIVE LOGITS

ipedia

1.25

idge

1.23

nik

1.17

ⓘ

1.15

geist

1.14

cker

1.12

DragonMagazine

1.09

atu

1.04

idad

1.03

č

1.01

Activations Density 0.007%

repeated use of the word "Polis" and variations of "ikarpov" alongside other related terms

No Comments

No Known Activations