INDEX

Explanations

names or terms related to "Nak" and potentially "MotoGP"

oai_token-act-pair · gpt-3.5-turbo

names of individuals associated with prominent events or developments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Morty

-0.77

 Chargers

-0.77

Dynamic

-0.70

 Titans

-0.70

 restraints

-0.69

force

-0.68

ulf

-0.66

ravis

-0.66

¿

-0.65

ivity

-0.64

POSITIVE LOGITS

Nak

2.95

Mog

1.81

Bir

1.80

Bir

1.40

Kas

1.33

 Shim

1.28

Miy

1.24

 Schne

1.21

Mam

1.16

Git

1.08

Activations Density 0.021%

names or terms related to "Nak" and potentially "MotoGP"

names of individuals associated with prominent events or developments

No Comments

No Known Activations