INDEX

Explanations

terms related to Japanese animated television shows, characters, and magazines

oai_token-act-pair · gpt-3.5-turbo

references to specific media titles or franchises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

espie

-0.80

 Lyons

-0.79

 Slovakia

-0.71

 optics

-0.69

 prov

-0.68

 FedEx

-0.67

 Dixon

-0.67

 Trinidad

-0.67

CLA

-0.67

 benches

-0.66

POSITIVE LOGITS

etsu

1.42

ikuman

1.38

utsu

1.36

ugi

1.32

Åį

1.29

itsu

1.27

ichi

1.25

oku

1.25

atsu

1.24

uko

1.24

Activations Density 0.162%

terms related to Japanese animated television shows, characters, and magazines

references to specific media titles or franchises

No Comments

No Known Activations