INDEX

Explanations

instances of political discourse and explanations of policies or positions

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

0:0.02

1:0.03

2:0.03

3:0.04

4:0.06

5:0.14

6:0.02

7:0.04

8:0.02

9:0.45

10:0.02

11:0.09

Negative Logits

 reckoning

-2.25

lad

-2.12

dar

-1.97

 honoured

-1.94

abet

-1.84

endant

-1.81

 cent

-1.77

Mich

-1.74

 because

-1.73

itia

-1.69

POSITIVE LOGITS

 VIDEOS

2.79

WHY

2.77

why

2.60

why

2.59

ModLoader

2.35

 discrepancies

2.21

 inconsistencies

2.20

 proble

2.18

how

2.08

Physical

1.99

Activations Density 0.178%

instances of political discourse and explanations of policies or positions

No Known Activations