INDEX

Explanations

references to repeated statements or affirmations in discussions or reports

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

1:0.02

2:0.08

3:0.10

4:0.02

5:0.04

6:0.05

7:0.07

8:0.11

9:0.16

10:0.07

11:0.22

Negative Logits

ocene

-1.02

 anywhere

-1.02

ologist

-1.02

ebook

-1.00

 born

-0.99

 lived

-0.98

MIT

-0.96

ridge

-0.96

 situated

-0.96

Bin

-0.95

POSITIVE LOGITS

 horizont

1.33

 vigilance

1.31

 assurances

1.28

Weak

1.23

 srfAttach

1.22

 referen

1.19

DCS

1.16

 Password

1.15

 omission

1.14

Quote

1.14

Activations Density 0.010%

references to repeated statements or affirmations in discussions or reports

No Comments

No Known Activations