INDEX

Explanations

words related to legal documents and historical events

oai_token-act-pair · gpt-3.5-turbo

references to significant historical and legal terms or concepts, particularly related to rights and governance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

terday

-0.94

 cous

-0.66

Ô

-0.65

theless

-0.64

 beware

-0.64

 profiling

-0.63

itars

-0.63

rx

-0.62

 physic

-0.60

ãĤ½

-0.60

POSITIVE LOGITS

 Bureau

1.01

 Council

1.01

 Legislature

0.97

 Committee

0.97

 Commission

0.96

 Chamber

0.96

Era

0.95

 Continent

0.94

 Convention

0.94

 Department

0.92

Activations Density 0.275%

words related to legal documents and historical events

references to significant historical and legal terms or concepts, particularly related to rights and governance

No Comments

No Known Activations