INDEX

Explanations

constructs related to online information credibility and skepticism

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

0:0.11

1:0.08

2:0.07

3:0.04

4:0.04

5:0.03

6:0.06

7:0.26

8:0.03

9:0.04

10:0.16

11:0.04

Negative Logits

 sett

-2.58

 Witcher

-2.36

 SHAR

-2.28

ker

-2.28

 Hezbollah

-2.21

SIM

-2.19

�

-2.18

sim

-2.15

 stub

-2.14

 basket

-2.13

POSITIVE LOGITS

Sn

2.80

Gr

2.24

 Esper

2.09

Tre

2.07

 Bonds

2.06

 Ages

2.06

Sn

2.03

[+

2.03

 Expend

2.00

 extradition

1.96

Activations Density 0.003%

constructs related to online information credibility and skepticism

No Known Activations