INDEX

Explanations

terms related to something new or recently released

oai_token-act-pair · gpt-3.5-turbo

references to something being newly introduced or recently created

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

ĺħ

-0.87

 Citation

-0.72

ashtra

-0.71

rice

-0.69

 Waste

-0.68

 Grail

-0.67

halla

-0.67

aminer

-0.66

otaur

-0.65

retty

-0.65

POSITIVE LOGITS

wed

1.12

 appointed

0.99

mint

0.97

 renovated

0.96

 mint

0.95

 acquired

0.90

 RELE

0.90

 liberated

0.88

 installed

0.86

 renamed

0.85

Activations Density 0.025%

terms related to something new or recently released

references to something being newly introduced or recently created

No Comments

No Known Activations