INDEX

Explanations

names of people and organizations

oai_token-act-pair · gpt-3.5-turbo

instances of verbs that indicate action or involvement in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

govtrack

-0.64

dayName

-0.63

Sport

-0.63

estamp

-0.63

istine

-0.62

éĥ

-0.61

inventoryQuantity

-0.61

AME

-0.61

manac

-0.60

displayText

-0.59

POSITIVE LOGITS

 dozens

1.58

 numerous

1.44

 several

1.42

 countless

1.34

 nine

1.29

 hundreds

1.27

 innumerable

1.26

 thirteen

1.25

 eleven

1.25

 seven

1.23

Activations Density 0.643%

names of people and organizations

instances of verbs that indicate action or involvement in various contexts

No Comments

No Known Activations