INDEX

Explanations

references to occupations, particularly in the context of creative roles

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

0:0.14

1:0.10

2:0.04

3:0.09

4:0.05

5:0.02

6:0.13

7:0.08

8:0.06

9:0.02

10:0.05

11:0.17

Negative Logits

 emer

-2.85

Pok

-2.64

 Bram

-2.63

FTWARE

-2.52

 jewelry

-2.37

 Akron

-2.33

 Coco

-2.32

 Cologne

-2.30

CI

-2.28

ophy

-2.28

POSITIVE LOGITS

 caption

5.88

 Caption

3.87

 subtitle

3.82

 capt

3.71

 subtitles

3.20

TIT

3.00

Capt

2.89

 IMAGES

2.79

rss

2.74

img

2.69

Activations Density 0.000%

references to occupations, particularly in the context of creative roles

No Known Activations