INDEX

Explanations

references to a specific automobile brand, Honda

oai_token-act-pair · gpt-3.5-turbo

references to Honda vehicles and their models

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ablishment

-0.94

asure

-0.89

ockets

-0.81

byter

-0.78

owder

-0.77

terday

-0.77

vironment

-0.76

onyms

-0.73

acular

-0.73

ools

-0.72

POSITIVE LOGITS

 Honda

1.16

 Civic

1.05

 Accord

1.00

 Mazda

0.83

 Yamaha

0.81

asaki

0.80

 Toyota

0.76

onda

0.76

bike

0.74

 Karin

0.74

Activations Density 0.010%

references to a specific automobile brand, Honda

references to Honda vehicles and their models

No Comments

No Known Activations