INDEX

Explanations

phrases and terms associated with rewards and incentives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.13

1:0.34

2:0.03

3:0.03

4:0.02

5:0.15

6:0.02

7:0.01

8:0.08

9:0.07

10:0.03

11:0.02

Negative Logits

ibu

-1.55

 Goth

-1.52

Hun

-1.46

 Anarchy

-1.43

atform

-1.42

Ada

-1.42

 Barbarian

-1.41

Raw

-1.40

tf

-1.39

Bug

-1.37

POSITIVE LOGITS

quet

1.85

 will

1.79

 continu

1.79

DonaldTrump

1.78

 isEnabled

1.68

onte

1.66

ucl

1.65

Lago

1.63

 should

1.63

will

1.62

Activations Density 0.024%

phrases and terms associated with rewards and incentives

No Comments

No Known Activations