INDEX

Explanations

terms related to honesty, trustworthiness, and ethics

oai_token-act-pair · gpt-3.5-turbo

references to the concept of integrity 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 @ 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Jet

-0.73

Stock

-0.71

bus

-0.70

GS

-0.67

--------------------------------------------------------

-0.66

Lear

-0.66

arters

-0.65

 Adventures

-0.65

Vert

-0.64

opa

-0.64

POSITIVE LOGITS

 integrity

1.49

rity

1.22

 Integrity

1.17

ulence

0.90

uala

0.85

orously

0.84

teness

0.84

 preservation

0.84

 credentials

0.83

acies

0.82

Activations Density 0.006%

terms related to honesty, trustworthiness, and ethics

references to the concept of integrity in various contexts

No Comments

No Known Activations