INDEX

Explanations

keywords related to influence or influencers

oai_token-act-pair · gpt-3.5-turbo

terms related to influencers and influence in social contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 whisk

-0.70

Dep

-0.69

 abst

-0.68

Rat

-0.66

 Plane

-0.65

Log

-0.63

 Kats

-0.62

Bus

-0.62

 alloc

-0.61

 Depend

-0.59

POSITIVE LOGITS

encer

4.60

encers

4.40

enced

2.37

encing

2.13

ences

2.07

ENC

1.87

enza

1.85

encia

1.75

enc

1.73

ence

1.68

Activations Density 0.014%

keywords related to influence or influencers

terms related to influencers and influence in social contexts

No Comments

No Known Activations