INDEX

Explanations

mentions of militant groups or individuals associated with terrorism

oai_token-act-pair · gpt-3.5-turbo

references to militant groups and their activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

IENCE

-0.84

othy

-0.80

ðĿ

-0.76

è¯

-0.75

æ©Ł

-0.74

encers

-0.73

ORK

-0.71

Doctor

-0.71

 Revel

-0.69

 Agric

-0.68

POSITIVE LOGITS

 militant

0.93

 militants

0.87

 Islamist

0.87

 Islamists

0.87

 insurgency

0.84

ilitary

0.83

 affiliated

0.80

Enlarge

0.80

anamo

0.77

 gunmen

0.76

Activations Density 0.029%

mentions of militant groups or individuals associated with terrorism

references to militant groups and their activities

No Comments

No Known Activations