INDEX

Explanations

instances of alleged wrongdoing or misconduct within various contexts

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to allegations or instances of wrongdoing and misconduct

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-att-kk

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

0:0.06

1:0.05

2:0.12

3:0.06

4:0.09

5:0.10

6:0.03

7:0.04

8:0.08

9:0.21

10:0.08

11:0.03

Negative Logits

 glide

-1.56

 todd

-1.37

soDeliveryDate

-1.36

 inhabited

-1.35

版

-1.35

amorph

-1.31

env

-1.30

rak

-1.25

anty

-1.22

veland

-1.22

POSITIVE LOGITS

 misconduct

1.76

 negligence

1.49

 wrongdoing

1.46

 inaction

1.45

 unlawful

1.44

 wrongful

1.43

 improper

1.43

 allegations

1.42

 Scientology

1.39

 Conduct

1.39

Activations Density 0.007%

instances of alleged wrongdoing or misconduct within various contexts

references to allegations or instances of wrongdoing and misconduct

No Comments

No Known Activations