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

1:0.09

2:0.07

3:0.08

4:0.08

5:0.07

6:0.08

7:0.08

8:0.08

9:0.07

10:0.08

11:0.08

Negative Logits

 Received

-1.90

HIM

-1.71

luaj

-1.64

 uncovered

-1.59

 Cursed

-1.59

 herself

-1.58

gettable

-1.54

emort

-1.53

 hers

-1.52

 Internal

-1.51

POSITIVE LOGITS

natureconservancy

1.82

 simultane

1.66

 preaching

1.55

 intens

1.55

\-

1.53

 Commando

1.52

 supplementation

1.52

 Apostles

1.52

 Apostle

1.51

 additives

1.50

Activations Density 0.000%

No Known Activations

This feature has no known activations.