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

1:0.05

2:0.10

3:0.09

4:0.08

5:0.07

6:0.08

7:0.07

8:0.07

9:0.07

10:0.07

11:0.10

Negative Logits

 Bris

-1.70

 Monitor

-1.60

Mix

-1.51

 agre

-1.48

 habitable

-1.48

 Revel

-1.40

Ref

-1.39

 Libre

-1.39

 Labor

-1.36

Cent

-1.36

POSITIVE LOGITS

tro

1.92

amy

1.86

bara

1.64

ucket

1.58

 admon

1.55

��

1.54

ikan

1.53

ruit

1.49

覚醒

1.46

verb

1.45

Activations Density 0.000%

No Known Activations

This feature has no known activations.