INDEX

Explanations

mentions of political cabinets

oai_token-act-pair · gpt-3.5-turbo

references to government cabinets

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

owicz

-0.79

amples

-0.66

RED

-0.65

ighth

-0.63

earned

-0.61

American

-0.61

olver

-0.61

Ïģ

-0.60

icago

-0.59

POSITIVE LOGITS

 cabinet

1.32

 cabinets

1.14

 ministers

1.04

 secretaries

1.03

 minister

0.97

 Cabinet

0.96

 Ministers

0.86

 appointments

0.85

 ministry

0.85

 ministerial

0.84

Activations Density 0.005%

mentions of political cabinets

references to government cabinets

No Comments

No Known Activations