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

0:0.07

1:0.09

2:0.08

3:0.08

4:0.07

5:0.08

6:0.08

7:0.07

8:0.09

9:0.08

10:0.08

11:0.07

Negative Logits

Jackson

-2.84

Cand

-2.76

Blake

-2.73

esc

-2.67

ounce

-2.56

uese

-2.55

webkit

-2.55

 Chan

-2.52

Favorite

-2.52

Cache

-2.50

POSITIVE LOGITS

 neutron

2.85

 uranium

2.80

Mit

2.53

 NATO

2.53

 srfAttach

2.38

 Worse

2.38

 nuclear

2.36

 disadvant

2.33

 understatement

2.32

 redef

2.32

Activations Density 0.000%

No Known Activations

This feature has no known activations.