INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L7_Hcat_z_lr1.20e-03_l11.10e+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.7.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

1:0.07

2:0.08

3:0.08

4:0.07

5:0.08

6:0.08

7:0.08

8:0.08

9:0.07

10:0.09

11:0.09

Negative Logits

 Oper

-1.96

 Orchestra

-1.93

 Gallery

-1.92

 conve

-1.86

 weaving

-1.86

 archives

-1.85

 Paid

-1.79

 Rove

-1.78

 Productions

-1.77

 Franchise

-1.76

POSITIVE LOGITS

 unpop

2.18

ratulations

2.13

ipel

2.09

anmar

2.06

acid

2.02

prus

2.02

nesota

1.99

cit

1.99

anooga

1.97

iversal

1.93

Activations Density 0.000%

No Known Activations

This feature has no known activations.