INDEX

Explanations

No Explanations Found

New Auto-Interp

Top 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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.06

1:0.08

2:0.08

3:0.08

4:0.10

5:0.08

6:0.08

7:0.08

8:0.07

9:0.08

10:0.08

11:0.07

Negative Logits

 copied

-1.51

 Sting

-1.48

 loyal

-1.43

 Riding

-1.42

style

-1.42

 silenced

-1.40

�

-1.39

 defeating

-1.39

 outright

-1.39

 Received

-1.37

POSITIVE LOGITS

itaire

2.28

raq

2.04

lde

1.91

oglu

1.89

️

1.82

orem

1.77

uberty

1.77

ukong

1.77

uala

1.73

gerald

1.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.