INDEX

Explanations

references to specific food items and brands, particularly in relation to pizza and fast food

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

0:0.04

1:0.03

2:0.09

3:0.19

4:0.03

5:0.03

6:0.13

7:0.22

8:0.03

9:0.03

10:0.08

11:0.05

Negative Logits

owship

-1.44

ottesville

-1.35

ngth

-1.30

 Lauder

-1.26

 bond

-1.24

 observer

-1.18

orescent

-1.15

 Moody

-1.11

 Lands

-1.08

uyomi

-1.08

POSITIVE LOGITS

ppers

1.25

nuts

1.16

emonium

1.10

 Hobby

1.07

 Puzz

1.06

 Cafe

1.05

ttes

1.04

iasco

1.04

Sweet

1.04

 pasta

1.02

Activations Density 0.016%

references to specific food items and brands, particularly in relation to pizza and fast food

No Known Activations