INDEX

Explanations

phrases or sentences indicating knowledge or awareness of certain information

oai_token-act-pair · gpt-3.5-turbo

references to levels of awareness or knowledge about specific threats or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ontent

-0.81

soever

-0.74

isma

-0.73

UME

-0.73

archy

-0.69

quer

-0.67

erer

-0.63

hement

-0.62

iment

-0.61

CG

-0.61

POSITIVE LOGITS

 impending

1.16

 irregularities

0.83

 wrongdoing

0.83

 dangers

0.81

 plans

0.79

 developments

0.78

how

0.77

 discrepancies

0.77

 injust

0.75

 pending

0.74

Activations Density 0.163%

phrases or sentences indicating knowledge or awareness of certain information

references to levels of awareness or knowledge about specific threats or events

No Comments

No Known Activations