INDEX

Explanations

terms related to acts of terrorism

oai_token-act-pair · gpt-3.5-turbo

references to terrorism and its impacts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

resso

-0.71

 Quartz

-0.69

flush

-0.68

endum

-0.68

galitarian

-0.68

bye

-0.65

dit

-0.63

 Grape

-0.62

rovers

-0.62

 Decker

-0.60

POSITIVE LOGITS

fully

1.04

fulness

0.92

izing

0.89

ously

0.86

 attacks

0.86

efully

0.85

 tactics

0.83

wave

0.82

abad

0.81

 raids

0.81

Activations Density 0.013%

terms related to acts of terrorism

references to terrorism and its impacts

No Comments

No Known Activations