INDEX

Explanations

elements related to inclusion and participation in specific contexts

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.01

2:0.13

3:0.07

4:0.08

5:0.04

6:0.07

7:0.31

8:0.03

9:0.05

10:0.08

11:0.06

Negative Logits

realDonaldTrump

-1.71

MODE

-1.65

quickShipAvailable

-1.61

 independ

-1.55

 behavi

-1.53

urred

-1.51

 recomm

-1.50

 independently

-1.49

 relent

-1.48

oward

-1.48

POSITIVE LOGITS

atto

1.57

atel

1.55

 boxes

1.53

ppings

1.52

TI

1.51

 lists

1.49

 Exhibit

1.47

 Schedule

1.46

 Gleaming

1.45

 Appendix

1.44

Activations Density 0.002%

elements related to inclusion and participation in specific contexts

No Comments

No Known Activations