INDEX

Explanations

references to plans, strategies, or blueprints

oai_token-act-pair · gpt-3.5-turbo

references to strategic plans or frameworks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Fi

-0.78

Viv

-0.73

sr

-0.71

Sn

-0.70

Sue

-0.69

 disclaim

-0.66

 Athe

-0.65

 showers

-0.65

Qu

-0.63

Sor

-0.63

POSITIVE LOGITS

 blueprint

3.47

 Blueprint

2.26

 roadmap

1.87

 timetable

1.66

 ledger

1.62

 manifesto

1.55

 playbook

1.48

 schematic

1.36

 precursor

1.23

 plan

1.20

Activations Density 0.016%

references to plans, strategies, or blueprints

references to strategic plans or frameworks

No Comments

No Known Activations