INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L7_Hcat_z_lr1.20e-03_l11.10e+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.7.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.05

1:0.08

2:0.21

3:0.06

4:0.03

5:0.06

6:0.05

7:0.10

8:0.04

9:0.05

10:0.15

11:0.07

Negative Logits

�

-2.68

SHA

-2.53

apor

-2.38

 Fool

-2.31

CLA

-2.21

�

-2.18

 Haram

-2.12

ikarp

-2.11

Lur

-2.07

crim

-2.06

POSITIVE LOGITS

ortment

2.66

 outings

2.65

 fixtures

2.52

 journeys

2.48

 Weeks

2.43

 Judd

2.42

 inspections

2.33

runs

2.30

 withdrawals

2.24

 slate

2.19

Activations Density 0.000%

No Known Activations

This feature has no known activations.