INDEX

Explanations

references to dangerous or lethal situations

oai_token-act-pair · gpt-3.5-turbo

references to fatal incidents or hazardous situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ourced

-0.88

agate

-0.83

ership

-0.82

arte

-0.81

erous

-0.81

arity

-0.81

orers

-0.78

estamp

-0.78

 CLIENT

-0.77

yk

-0.76

POSITIVE LOGITS

 poisonous

0.94

 wounding

0.91

 poison

0.90

 deadly

0.89

 assault

0.85

 lethal

0.82

 dose

0.81

 toll

0.81

 fighting

0.78

 overdose

0.78

Activations Density 0.010%

references to dangerous or lethal situations

references to fatal incidents or hazardous situations

No Comments

No Known Activations