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

Head Attr Weights

0:0.09

1:0.08

2:0.07

3:0.07

4:0.08

5:0.08

6:0.07

7:0.07

8:0.09

9:0.08

10:0.09

11:0.08

Negative Logits

yrinth

-1.92

eto

-1.49

quished

-1.44

asionally

-1.43

��

-1.43

 cruel

-1.42

udget

-1.42

uel

-1.41

inge

-1.40

urst

-1.39

POSITIVE LOGITS

 SEAL

1.53

 Appl

1.49

 Newsp

1.48

 inhal

1.44

 Scand

1.41

 analogy

1.41

 insulation

1.41

 withstand

1.39

 smell

1.39

 scales

1.38

Activations Density 0.000%

No Known Activations

This feature has no known activations.