INDEX

Explanations

references to the publication "The Nation" and actions associated with its content, such as subscribing or checking for new issues

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

Head Attr Weights

0:0.02

1:0.07

2:0.08

3:0.07

4:0.04

5:0.16

6:0.08

7:0.07

8:0.08

9:0.08

10:0.07

11:0.12

Negative Logits

nikov

-1.64

EVA

-1.43

(?,

-1.40

utenant

-1.32

edi

-1.29

urized

-1.28

enko

-1.26

 additives

-1.26

ovsky

-1.25

affer

-1.22

POSITIVE LOGITS

iphate

1.68

 redistribution

1.43

ModLoader

1.42

 welf

1.41

 intrusive

1.41

Amazon

1.37

headers

1.35

Pixel

1.35

Swe

1.33

roud

1.32

Activations Density 0.004%

references to the publication "The Nation" and actions associated with its content, such as subscribing or checking for new issues

No Known Activations