INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.14

2:0.06

3:0.08

4:0.06

5:0.06

6:0.12

7:0.07

8:0.08

9:0.07

10:0.08

11:0.05

Negative Logits

ificate

-1.53

 Remix

-1.47

 remix

-1.45

iverse

-1.44

lyr

-1.41

umbs

-1.40

 danced

-1.40

Woo

-1.39

acular

-1.38

 queens

-1.37

POSITIVE LOGITS

士

1.76

 Ammunition

1.58

uay

1.56

nance

1.50

cair

1.45

aq

1.43

urat

1.43

onnaissance

1.43

uminium

1.40

recated

1.38

Activations Density 0.000%

No Known Activations

This feature has no known activations.