INDEX

Explanations

words related to official positions or titles

oai_token-act-pair · gpt-3.5-turbo

references to official positions or roles in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

¿½

-0.78

 phosphate

-0.76

 breath

-0.68

 fertility

-0.63

 uranium

-0.62

back

-0.61

 immersion

-0.61

FORM

-0.61

dding

-0.60

 breathing

-0.60

POSITIVE LOGITS

uously

1.20

eless

1.15

etr

1.15

uous

1.11

emate

1.10

ulatory

1.10

ements

1.09

ues

1.06

orial

1.05

ework

1.03

Activations Density 0.045%

words related to official positions or titles

references to official positions or roles in various contexts

No Comments

No Known Activations