INDEX

Explanations

descriptions of past experiences or fictional scenarios

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

phrases that indicate a reason or explanation

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

2:0.08

3:0.04

4:0.02

5:0.05

6:0.06

7:0.05

8:0.05

9:0.05

10:0.05

11:0.46

Negative Logits

lished

-1.33

 datas

-1.27

bye

-1.22

 butcher

-1.21

 standings

-1.21

ドラ

-1.19

 vict

-1.18

��

-1.15

mons

-1.13

POSITIVE LOGITS

athom

1.20

 overshadowed

1.19

oil

1.18

 subconscious

1.17

shared

1.16

Area

1.16

NetMessage

1.16

 Canal

1.15

onite

1.13

UV

1.12

Activations Density 0.113%

descriptions of past experiences or fictional scenarios

phrases that indicate a reason or explanation

No Known Activations

descriptions of past experiences or fictional scenarios

phrases that indicate a reason or explanation

No Known Activations