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

0:0.08

1:0.06

2:0.08

3:0.08

4:0.07

5:0.08

6:0.07

7:0.07

8:0.10

9:0.09

10:0.08

11:0.08

Negative Logits

vice

-1.87

 reperto

-1.74

yp

-1.68

quer

-1.65

opic

-1.61

hell

-1.53

major

-1.53

 cylinders

-1.48

(_

-1.46

osterone

-1.46

POSITIVE LOGITS

geoning

2.04

 Dhabi

1.70

raq

1.69

 RAND

1.60

Yug

1.59

 admission

1.54

 Gabriel

1.53

 artisan

1.53

ideon

1.52

 dusk

1.48

Activations Density 0.000%

No Known Activations

This feature has no known activations.