INDEX

Explanations

Phrases related to legislative measures or bills introduced for public vote

oai_token-act-pair · gpt-3.5-turbo

mentions of legislative measures or bills

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

 Continent

-0.82

 Expedition

-0.72

elsen

-0.69

igi

-0.69

Naz

-0.67

bledon

-0.64

 Assassins

-0.63

 Canterbury

-0.62

elong

-0.62

 communal

-0.62

POSITIVE LOGITS

 legalizing

1.04

 authorizing

0.94

 decriminal

0.88

 proposing

0.82

HB

0.79

 repealing

0.78

 unanimously

0.77

ospons

0.75

 prohibiting

0.72

 vetoed

0.72

Activations Density 0.081%

Phrases related to legislative measures or bills introduced for public vote

mentions of legislative measures or bills

No Comments

No Known Activations