INDEX

Explanations

news-related events or headlines involving violence or conflict

oai_token-act-pair · gpt-3.5-turbo

phrases related to violent incidents or threats

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

ichever

-0.78

 gradual

-0.65

 coales

-0.64

omever

-0.63

 increment

-0.62

respective

-0.61

erning

-0.61

 cumulative

-0.60

icip

-0.60

 ideally

-0.60

POSITIVE LOGITS

 Saying

0.91

 Posted

0.90

 Copyright

0.89

pic

0.88

 raping

0.86

bc

0.86

 Belfast

0.83

ãħĭãħĭ

0.83

↵Âł

0.82

Â¶

0.81

Activations Density 0.628%

news-related events or headlines involving violence or conflict

phrases related to violent incidents or threats

No Comments

No Known Activations