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

1:0.11

2:0.08

3:0.07

4:0.07

5:0.09

6:0.08

7:0.07

8:0.07

9:0.08

10:0.08

11:0.08

Negative Logits

 Style

-1.64

 Press

-1.60

ulator

-1.59

riter

-1.59

 Dell

-1.59

 Huawei

-1.58

 BlackBerry

-1.56

 Draper

-1.56

QC

-1.56

 Bern

-1.55

POSITIVE LOGITS

atl

1.78

gomery

1.75

 veter

1.67

 obser

1.58

 landfall

1.57

 adverse

1.55

 withd

1.55

 retake

1.49

 administer

1.49

ocent

1.49

Activations Density 0.000%

No Known Activations

This feature has no known activations.