INDEX

Explanations

references to violent and gruesome acts, including executions, mutilation, and mass killings

oai_token-act-pair · gpt-3.5-turbo

references to acts of violence and brutality

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nect

-0.90

 Compass

-0.87

ometer

-0.83

 Startup

-0.83

 Transition

-0.83

Premium

-0.82

izons

-0.80

 Outlook

-0.78

 connectivity

-0.78

 Clouds

-0.78

POSITIVE LOGITS

 murderers

1.91

 murdering

1.86

 murders

1.85

 killings

1.83

 murder

1.80

 massacres

1.78

 slaughter

1.78

 murderer

1.73

 atrocities

1.73

 killers

1.71

Activations Density 1.022%

references to violent and gruesome acts, including executions, mutilation, and mass killings

references to acts of violence and brutality

No Comments

No Known Activations