INDEX

Explanations

proper nouns representing individuals or entities that have been approved or selected by a particular authority

oai_token-act-pair · gpt-3.5-turbo

phrases indicating authorship or attribution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ettes

-0.73

itiveness

-0.72

iasm

-0.70

heimer

-0.70

itled

-0.70

ivia

-0.70

aunder

-0.69

 largeDownload

-0.69

pect

-0.68

clips

-0.68

POSITIVE LOGITS

 virtue

1.06

 decree

0.88

 statute

0.84

product

0.80

products

0.79

 default

0.76

 parliament

0.75

 fiat

0.75

 invitation

0.74

laws

0.71

Activations Density 0.164%

proper nouns representing individuals or entities that have been approved or selected by a particular authority

phrases indicating authorship or attribution

No Comments

No Known Activations