INDEX

Explanations

quotes from individuals discussing various topics and their opinions

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

0:0.03

1:0.36

2:0.05

3:0.02

4:0.05

5:0.14

6:0.05

7:0.04

8:0.04

9:0.06

10:0.07

11:0.04

Negative Logits

!'

-2.01

DragonMagazine

-1.91

?'

-1.89

 transformer

-1.77

 concess

-1.69

'?

-1.68

zbollah

-1.66

hovah

-1.60

,'

-1.60

usercontent

-1.57

POSITIVE LOGITS

Said

2.44

 "...

1.85

 quoted

1.76

ヴ

1.71

Desc

1.71

 Aval

1.70

 Said

1.68

said

1.60

 Cummings

1.60

Adds

1.60

Activations Density 0.278%

quotes from individuals discussing various topics and their opinions

No Known Activations