INDEX

Explanations

mentions of officials or people in authoritative positions

oai_token-act-pair · gpt-3.5-turbo

mentions of officials or authority figures

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

Beg

-0.84

ï¸

-0.80

Õ

-0.71

âĶģ

-0.71

âĶĢâĶĢâĶĢâĶĢ

-0.70

 Bengal

-0.69

esville

-0.68

Interstitial

-0.67

DAY

-0.65

THER

-0.65

POSITIVE LOGITS

 officials

0.88

 overseeing

0.87

assetsadobe

0.84

 tasked

0.81

ially

0.80

hips

0.80

iets

0.77

 Officials

0.77

IAL

0.74

 stationed

0.73

Activations Density 0.037%

mentions of officials or people in authoritative positions

mentions of officials or authority figures

No Comments

No Known Activations