INDEX

Explanations

first-person pronouns and various forms of the verb 'to be'

oai_token-act-pair · gpt-3.5-turbo

personal pronouns and expressions of personal opinion or beliefs

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

Raven

-0.46

 Tooth

-0.44

 Hogwarts

-0.43

 Sirius

-0.43

 Poke

-0.43

ModLoader

-0.43

-0.42

Offline

-0.41

 Sparrow

-0.41

 Scouting

-0.41

POSITIVE LOGITS

enegger

0.60

allery

0.57

umenthal

0.53

aspers

0.51

rored

0.51

haus

0.51

achev

0.50

zbek

0.50

ividual

0.49

vironment

0.49

Activations Density 2.892%

first-person pronouns and various forms of the verb 'to be'

personal pronouns and expressions of personal opinion or beliefs

No Known Activations