INDEX

Explanations

expressions of uncertainty and self-reflection

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

0:0.02

1:0.09

2:0.12

3:0.04

4:0.02

5:0.03

6:0.20

7:0.10

8:0.04

9:0.04

10:0.08

11:0.18

Negative Logits

osate

-1.37

 Breaker

-1.23

 Sour

-1.17

SPI

-1.11

 horm

-1.11

nown

-1.11

Mat

-1.10

 Melt

-1.10

 spiked

-1.09

Esp

-1.07

POSITIVE LOGITS

hunt

1.37

bernatorial

1.35

arial

1.29

soever

1.27

ele

1.24

rous

1.24

Synopsis

1.24

HB

1.23

enery

1.18

bern

1.18

Activations Density 0.022%

expressions of uncertainty and self-reflection

No Known Activations