INDEX

Explanations

mentions of military ranks or units in the context of an organization or structure

oai_token-act-pair · gpt-3.5-turbo

mentions of military units or ranks

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

 Gutenberg

-0.73

 Ceres

-0.67

 Franken

-0.67

 Realms

-0.67

payer

-0.67

FDA

-0.66

 Hungarian

-0.62

onomy

-0.62

 Gingrich

-0.62

BUS

-0.61

POSITIVE LOGITS

adier

1.45

itte

1.09

 brig

1.05

aded

0.99

gae

0.98

ments

0.97

ading

0.96

antine

0.96

alions

0.95

isons

0.93

Activations Density 0.019%

mentions of military ranks or units in the context of an organization or structure

mentions of military units or ranks

No Comments

No Known Activations