INDEX

Explanations

dates written in a specific format

oai_token-act-pair · gpt-3.5-turbo

dates and numbers associated with news articles

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

 Brawl

-0.86

OU

-0.85

 Clause

-0.82

 Fisheries

-0.80

 Clinic

-0.79

LTD

-0.78

 Thrones

-0.77

 Zone

-0.77

 Covenant

-0.76

EMS

-0.74

POSITIVE LOGITS

hillary

1.61

why

1.55

john

1.45

trump

1.45

united

1.41

christ

1.40

how

1.39

milo

1.38

exclusive

1.38

maxwell

1.36

Activations Density 0.040%

dates written in a specific format

dates and numbers associated with news articles

No Comments

No Known Activations