INDEX

Explanations

commands or actions related to problem-solving or making decisions

New Auto-Interp

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

1:0.01

2:0.10

3:0.05

4:0.12

5:0.03

6:0.05

7:0.24

8:0.05

9:0.03

10:0.13

11:0.11

Negative Logits

 ORIG

-1.30

 revision

-1.27

 murderers

-1.22

 referenced

-1.21

ants

-1.21

 Revision

-1.20

ciation

-1.20

 originated

-1.20

origin

-1.17

pim

-1.17

POSITIVE LOGITS

cffff

1.62

iolet

1.43

uno

1.37

ierrez

1.36

ntil

1.35

 Bless

1.35

hower

1.27

oustic

1.24

veland

1.24

NRS

1.23

Activations Density 0.012%

commands or actions related to problem-solving or making decisions

No Known Activations