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

0:0.08

1:0.04

2:0.09

3:0.08

4:0.09

5:0.09

6:0.08

7:0.07

8:0.08

9:0.09

10:0.08

11:0.08

Negative Logits

sonian

-1.80

abba

-1.56

 rake

-1.51

 mound

-1.46

izzard

-1.46

 helicop

-1.45

──

-1.40

 Rolls

-1.40

 Airbnb

-1.39

 toilet

-1.37

POSITIVE LOGITS

VID

1.82

RFC

1.65

ilver

1.62

QU

1.62

Supplement

1.53

武

1.49

 Vance

1.47

ults

1.46

articles

1.45

neutral

1.45

Activations Density 0.000%

No Known Activations

This feature has no known activations.