INDEX

Explanations

references to the Islamic State group

oai_token-act-pair · gpt-3.5-turbo

mentions of the Islamic State

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Canal

-0.77

 Dent

-0.70

 nutrit

-0.69

 Turing

-0.69

neau

-0.68

 Macy

-0.67

quin

-0.66

isher

-0.65

ishers

-0.65

iddy

-0.64

POSITIVE LOGITS

 Boko

1.01

 Hezbollah

0.97

Nusra

0.97

 Donetsk

0.94

 paramilitary

0.93

ullah

0.92

 insurgency

0.91

iannopoulos

0.88

 militant

0.87

 militants

0.86

Activations Density 0.548%

references to the Islamic State group

mentions of the Islamic State

No Comments

No Known Activations