INDEX

Explanations

terms related to security and management concepts

New Auto-Interp

Features by Cosine Similarity

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

Plots

Explanation

Show Test Field

Default Test Text

IFrame

Link

Not in Any Lists

Head Attr Weights

0:0.03

1:0.02

2:0.01

3:0.02

4:0.02

5:0.62

6:0.02

7:0.01

8:0.04

9:0.08

10:0.05

11:0.03

Negative Logits

 sonic

-1.50

obe

-1.48

Mug

-1.45

 constitu

-1.44

Balt

-1.38

 Sleeping

-1.36

-1.34

Rock

-1.33

 Negro

-1.32

Tup

-1.32

POSITIVE LOGITS

and

2.18

2.10

バ

1.74

ands

1.69

AND

1.63

CTR

1.61

±

1.61

1.60

or

1.60

ihara

1.58

Activations Density 0.711%

terms related to security and management concepts

No Activations Uploaded