INDEX

Explanations

names of extremist groups

oai_token-act-pair · gpt-3.5-turbo

references to militant or terrorist groups

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

othy

-0.72

Labor

-0.71

è¯

-0.70

 Shore

-0.69

 pollen

-0.68

 Lect

-0.67

vent

-0.65

 Wheat

-0.64

ðĿ

-0.64

Consumer

-0.63

POSITIVE LOGITS

 Boko

0.95

wana

0.94

affiliated

0.88

 affiliation

0.84

atics

0.84

 affili

0.83

abul

0.83

alions

0.83

 insurgents

0.82

 jihadists

0.81

Activations Density 0.044%

names of extremist groups

references to militant or terrorist groups

No Comments

No Known Activations