INDEX

Explanations

phrases related to incentives or motivations

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize the effects and implications of actions or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

untled

-0.79

actionDate

-0.75

ominated

-0.72

anmar

-0.71

rue

-0.71

ItemThumbnailImage

-0.70

 Archdemon

-0.70

 descended

-0.66

numbered

-0.66

ailability

-0.65

POSITIVE LOGITS

 helps

1.74

 enhances

1.62

 improves

1.59

 strengthens

1.58

 boosts

1.55

 enables

1.53

 reduces

1.47

 protects

1.47

 saves

1.47

 allows

1.47

Activations Density 0.467%

phrases related to incentives or motivations

phrases that emphasize the effects and implications of actions or behaviors

No Comments

No Known Activations