INDEX

Explanations

mentions of historical or real-world events and settings

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to reading and related activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-att-kk

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.02

2:0.21

3:0.07

4:0.09

5:0.05

6:0.02

7:0.01

8:0.22

9:0.14

10:0.04

11:0.02

Negative Logits

akable

-1.33

 porous

-1.23

xious

-1.04

ofi

-1.03

 confounding

-1.02

 easiest

-1.02

onomous

-1.01

ounced

-1.00

 negotiators

-0.98

atu

-0.97

POSITIVE LOGITS

DragonMagazine

1.65

 aloud

1.60

 VIDEOS

1.54

版

1.54

dayName

1.35

 Pastebin

1.29

oslav

1.26

 Digest

1.25

 Dictionary

1.24

papers

1.24

Activations Density 0.019%

mentions of historical or real-world events and settings

references to reading and related activities

No Comments

No Known Activations