INDEX

Explanations

terms related to government policies and positions of authority

oai_token-act-pair · gpt-3.5-turbo

terms related to power and authority in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

owship

-0.83

ifts

-0.76

oldemort

-0.76

ĻĤ

-0.76

arb

-0.71

ILY

-0.70

iannopoulos

-0.67

perature

-0.67

livion

-0.66

ä¸Ĭ

-0.66

POSITIVE LOGITS

 Responsibility

1.10

 Accountability

1.00

 Manifest

0.97

Era

0.96

 Package

0.95

 Works

0.95

 Clause

0.95

 Authority

0.94

 Solution

0.94

 Guys

0.94

Activations Density 0.116%

terms related to government policies and positions of authority

terms related to power and authority in various contexts

No Comments

No Known Activations