INDEX

Explanations

phrases related to affiliation or association with individuals or groups

oai_token-act-pair · gpt-3.5-turbo

references to affiliations or connections to organizations

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

 Passage

-0.79

ettel

-0.70

bum

-0.66

erva

-0.65

perfect

-0.65

 warmth

-0.64

 Rousse

-0.64

lift

-0.63

 Vanity

-0.63

val

-0.61

POSITIVE LOGITS

 affiliated

3.37

affiliated

2.55

 affiliation

2.28

 affili

2.15

 unaff

2.11

 affiliates

2.06

iliated

2.05

 affiliate

2.03

 accredited

1.36

 allied

1.26

Activations Density 0.019%

phrases related to affiliation or association with individuals or groups

references to affiliations or connections to organizations

No Comments

No Known Activations