INDEX

Explanations

lists of actions or recommendations for readers

oai_token-act-pair · gpt-3.5-turbo

phrases indicating specific actions or activities that individuals can take regularly

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ilts

-0.72

 Unic

-0.67

 Blanc

-0.66

latest

-0.65

Alc

-0.64

oof

-0.63

cephal

-0.62

orf

-0.61

tymology

-0.61

Seb

-0.60

POSITIVE LOGITS

bleacher

0.80

APD

0.73

IVE

0.73

åĳ

0.71

æŃ¦

0.70

å°Ĩ

0.69

å¤©

0.68

 willpower

0.66

ðŁĳ

0.65

 karma

0.63

Activations Density 0.040%

lists of actions or recommendations for readers

phrases indicating specific actions or activities that individuals can take regularly

No Comments

No Known Activations