INDEX

Explanations

phrases related to apologies and accountability

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

New Auto-Interp

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

0:0.02

1:0.02

2:0.18

3:0.10

4:0.15

5:0.02

6:0.05

7:0.17

8:0.03

9:0.04

10:0.10

11:0.07

Negative Logits

aughtered

-1.52

TABLE

-1.51

acebook

-1.49

soDeliveryDate

-1.49

oother

-1.44

toggle

-1.43

ickr

-1.39

ーン

-1.35

oiler

-1.34

emale

-1.34

POSITIVE LOGITS

 whatsoever

1.86

 deem

1.50

 wrongdoing

1.44

soever

1.44

nor

1.42

 frivol

1.41

Discuss

1.41

 necessary

1.40

 warr

1.38

 mater

1.38

Activations Density 0.052%

phrases related to apologies and accountability

No Comments

No Known Activations