INDEX

Explanations

references to broadband or television-related terms

oai_token-act-pair · gpt-3.5-turbo

references to "Broad" or "Broadband" and related terminology

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Oro

-0.78

Downloadha

-0.73

 Uriel

-0.67

 Yanuk

-0.67

 Guerrero

-0.66

 Hugo

-0.65

 Canaver

-0.63

zzi

-0.63

Tup

-0.63

ICAN

-0.62

POSITIVE LOGITS

band

1.36

casters

1.35

sword

1.30

casts

1.28

casting

1.26

caster

1.25

bent

1.18

cast

1.18

sheet

1.12

cloth

1.11

Activations Density 0.035%

references to broadband or television-related terms

references to "Broad" or "Broadband" and related terminology

No Comments

No Known Activations