INDEX

Explanations

references to violent acts and crime

oai_token-act-pair · gpt-3.5-turbo

topics related to violence and human rights abuses

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

 Pebble

-0.87

 Canter

-0.85

ibilities

-0.81

irement

-0.75

 Flex

-0.74

Premium

-0.74

ellen

-0.74

hao

-0.73

Apple

-0.72

 Compass

-0.70

POSITIVE LOGITS

 massacres

1.84

 atroc

1.72

 massacre

1.70

 atrocities

1.66

 genocide

1.60

 killings

1.55

 slaughter

1.53

 murderers

1.51

 murders

1.51

 gruesome

1.50

Activations Density 1.078%

references to violent acts and crime

topics related to violence and human rights abuses

No Comments

No Known Activations