INDEX

Explanations

instances of significant numerical data or figures related to events

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

0:0.01

1:0.01

2:0.28

3:0.13

4:0.09

5:0.03

6:0.02

7:0.07

8:0.06

9:0.03

10:0.09

11:0.12

Negative Logits

ancies

-1.45

 accordingly

-1.42

ictionary

-1.42

ription

-1.41

 actionGroup

-1.38

 conscience

-1.35

ctica

-1.34

AE

-1.32

Vs

-1.31

Translation

-1.29

POSITIVE LOGITS

Oss

1.55

Osw

1.47

 Alam

1.42

 resumed

1.39

ocious

1.38

 fraught

1.34

 resur

1.33

 devoted

1.31

 reunited

1.30

bred

1.30

Activations Density 0.030%

instances of significant numerical data or figures related to events

No Known Activations