INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.9.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

0:0.10

1:0.08

2:0.08

3:0.08

4:0.06

5:0.06

6:0.08

7:0.07

8:0.08

9:0.08

10:0.08

11:0.09

Negative Logits

 Berks

-3.42

Sah

-2.89

 Reconstruction

-2.66

 Camden

-2.60

 Augusta

-2.58

 Berkeley

-2.52

Haw

-2.51

 Archae

-2.50

 theorem

-2.46

-2.45

POSITIVE LOGITS

ipment

2.65

 Trance

2.65

utic

2.63

uty

2.59

plet

2.58

acters

2.53

 Persona

2.48

 ratios

2.46

 Spirits

2.37

fx

2.35

Activations Density 0.000%

No Known Activations

This feature has no known activations.