INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.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

0:0.08

1:0.08

2:0.07

3:0.09

4:0.09

5:0.08

6:0.07

7:0.08

8:0.08

9:0.08

10:0.07

11:0.07

Negative Logits

 pony

-2.75

────

-2.74

 snail

-2.61

tube

-2.60

 barn

-2.54

.」

-2.52

 worms

-2.52

 reluctantly

-2.40

 wedd

-2.39

 patiently

-2.37

POSITIVE LOGITS

 Interior

2.71

 Cities

2.67

ivic

2.63

 Kaepernick

2.54

Park

2.52

 Basketball

2.42

 Schools

2.37

Edu

2.35

 Mental

2.33

Geh

2.31

Activations Density 0.000%

No Known Activations

This feature has no known activations.