INDEX

Explanations

mentions of government officials at the highest level, particularly individuals serving in the Cabinet

oai_token-act-pair · gpt-3.5-turbo

references to governmental cabinet positions and meetings

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

thia

-0.75

owicz

-0.71

imity

-0.70

ibaba

-0.67

Gim

-0.67

amples

-0.66

Ô

-0.66

 Wear

-0.64

izen

-0.64

 GOODMAN

-0.63

POSITIVE LOGITS

 secretaries

1.23

 ministers

1.13

 appointments

1.02

 minister

0.92

 Ministers

0.91

 secretary

0.90

chairs

0.87

 members

0.86

 nominees

0.85

 chairs

0.85

Activations Density 0.019%

mentions of government officials at the highest level, particularly individuals serving in the Cabinet

references to governmental cabinet positions and meetings

No Comments

No Known Activations