INDEX

Explanations

dates

oai_token-act-pair · gpt-3.5-turbo

references to individuals or names in a narrative context

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

NCT

-0.88

 monop

-0.83

 sidx

-0.82

ollah

-0.79

 nodd

-0.78

 influenza

-0.78

Hik

-0.78

 nursing

-0.76

accompan

-0.76

EStreamFrame

-0.75

POSITIVE LOGITS

 Stone

2.70

Stone

2.53

 Stones

2.07

 stone

2.01

 stones

1.84

stone

1.72

stones

1.59

 limestone

1.30

 granite

1.28

Glass

1.24

Activations Density 0.262%

dates

references to individuals or names in a narrative context

No Comments

No Known Activations