INDEX

Explanations

violent actions and weapons

oai_token-act-pair · gpt-3.5-turbo

references to violent impacts or forceful events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Tea

-0.82

soDeliveryDate

-0.80

parency

-0.79

 Waiting

-0.75

 Indigo

-0.74

 anonymity

-0.72

 Reserved

-0.71

acebook

-0.71

 Sched

-0.71

 confidentiality

-0.69

POSITIVE LOGITS

 blows

1.62

 punches

1.47

 blow

1.39

 punch

1.33

 hits

1.33

 blasts

1.32

 hammer

1.32

 projectiles

1.32

 hurled

1.31

 projectile

1.30

Activations Density 0.690%

violent actions and weapons

references to violent impacts or forceful events

No Comments

No Known Activations