INDEX

Explanations

phrases and terms associated with a specific count or quantity in various contexts

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

0:0.01

1:0.03

2:0.13

3:0.14

4:0.01

5:0.02

6:0.09

7:0.04

8:0.08

9:0.20

10:0.04

11:0.16

Negative Logits

worth

-1.40

acs

-1.31

istries

-1.28

 DRAG

-1.03

nces

-1.02

 unbeliev

-1.00

 responsiveness

-0.99

abases

-0.99

 temperament

-0.98

 legitimacy

-0.94

POSITIVE LOGITS

ereo

1.22

 province

1.14

 Donation

1.07

 Exclusive

1.07

"},{"

1.06

 Macedonia

1.03

 Activate

1.02

ortment

1.01

ulkan

1.00

EE

0.99

Activations Density 0.014%

phrases and terms associated with a specific count or quantity in various contexts

No Comments

No Known Activations