INDEX

Explanations

segments that contain temporal references or descriptions of events occurring over time

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_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.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.22

1:0.07

2:0.10

3:0.03

4:0.04

5:0.02

6:0.12

7:0.13

8:0.01

9:0.05

10:0.12

11:0.03

Negative Logits

Mexico

-3.28

 Mexicans

-3.25

Ec

-3.23

 Texans

-3.09

Mex

-3.04

 pokemon

-2.94

Tex

-2.91

 Luna

-2.90

Zap

-2.81

pes

-2.77

POSITIVE LOGITS

 Sweeney

9.34

weeney

4.87

Assembly

3.79

 Fletcher

3.44

MSG

3.37

 Walsh

3.24

 Hoffman

3.18

hire

3.14

 Worcester

3.10

ASP

3.09

Activations Density 0.000%

segments that contain temporal references or descriptions of events occurring over time

No Known Activations