INDEX

Explanations

mentions of militant groups and conflict areas

oai_token-act-pair · gpt-3.5-turbo

references to extremist groups and terrorism

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-1.07

bernatorial

-1.03

peak

-0.84

tesy

-0.84

abase

-0.82

ertation

-0.81

undrum

-0.79

ortality

-0.79

onse

-0.75

eanor

-0.75

POSITIVE LOGITS

 militants

1.52

 extremists

1.52

 forces

1.50

 militias

1.47

 insurgents

1.47

 groups

1.46

 gangs

1.46

 thugs

1.44

 separatists

1.43

 rebels

1.42

Activations Density 0.214%

mentions of militant groups and conflict areas

references to extremist groups and terrorism

No Comments

No Known Activations