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

0:0.11

1:0.07

2:0.08

3:0.08

4:0.07

5:0.08

6:0.06

7:0.07

8:0.09

9:0.09

10:0.07

11:0.06

Negative Logits

 Cheong

-1.79

 Merchants

-1.76

 vigilante

-1.76

lihood

-1.74

 Presbyter

-1.73

 enthusi

-1.71

 Presbyterian

-1.65

 pestic

-1.62

 Murd

-1.62

Esk

-1.58

POSITIVE LOGITS

quickShipAvailable

1.95

 diagram

1.88

initions

1.87

romy

1.83

uder

1.81

wcsstore

1.76

 mant

1.75

å

1.75

 rear

1.71

word

1.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.