INDEX

Explanations

phrases related to dominance and control

oai_token-act-pair · gpt-3.5-turbo

statements or advice related to interpersonal boundaries and conflict

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

çīĪ

-0.77

 toured

-0.76

 Census

-0.76

 Trends

-0.73

 Browse

-0.71

Maps

-0.71

 Bermuda

-0.70

 Trails

-0.69

phot

-0.69

 refurb

-0.66

POSITIVE LOGITS

 coercion

1.29

 coward

1.24

 retaliate

1.23

 coercive

1.19

 aggress

1.18

 cowardly

1.18

 coer

1.17

 disob

1.16

 provocation

1.16

 punish

1.15

Activations Density 0.723%

phrases related to dominance and control

statements or advice related to interpersonal boundaries and conflict

No Comments

No Known Activations