INDEX

Explanations

postpositions 'by' followed by a title and author

oai_token-act-pair · gpt-3.5-turbo

citations or credits indicating authorship in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ptions

-0.80

 redistributed

-0.74

ption

-0.73

});

-0.71

ptive

-0.71

cffffcc

-0.70

illance

-0.70

bia

-0.70

wcs

-0.69

bian

-0.69

POSITIVE LOGITS

 Edwin

1.04

 David

1.02

 Robert

1.01

 Colin

1.01

 Philip

1.01

 Jonathan

1.01

 Nathan

1.01

 William

1.00

 Joseph

1.00

 virtue

1.00

Activations Density 0.075%

postpositions 'by' followed by a title and author

citations or credits indicating authorship in a text

No Comments

No Known Activations