INDEX

Explanations

names ending in "zler" with varying levels of certainty or emphasis

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

words related to identification or categorization, particularly in the context of names or titles

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-att-kk

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

Head Attr Weights

0:0.02

1:0.03

2:0.05

3:0.38

4:0.02

5:0.02

6:0.06

7:0.13

8:0.04

9:0.06

10:0.06

11:0.06

Negative Logits

 IMAGES

-1.43

yahoo

-1.27

 VIDEOS

-1.25

mercial

-1.22

 Topics

-1.22

gmail

-1.16

MF

-1.14

ournals

-1.14

�

-1.13

irtual

-1.12

POSITIVE LOGITS

issance

1.70

isen

1.24

 worm

1.24

 ther

1.21

forced

1.21

standing

1.21

 laure

1.21

bal

1.19

tein

1.18

alore

1.18

Activations Density 0.006%

names ending in "zler" with varying levels of certainty or emphasis

words related to identification or categorization, particularly in the context of names or titles

No Known Activations

names ending in "zler" with varying levels of certainty or emphasis

words related to identification or categorization, particularly in the context of names or titles

No Known Activations