INDEX

Explanations

legal and legislative terms or actions

oai_token-act-pair · gpt-3.5-turbo

mentions of legislative bills and their specific provisions or effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

aths

-0.79

igans

-0.69

thia

-0.68

 coli

-0.67

 selves

-0.67

 adolesc

-0.66

 rooft

-0.65

 prolet

-0.65

 vantage

-0.64

Atl

-0.64

POSITIVE LOGITS

 unanimously

0.87

 Corker

0.86

 repealed

0.81

 amended

0.80

 Tillerson

0.77

 Fiscal

0.77

 lobbied

0.73

 jointly

0.73

 overseen

0.73

 McAuliffe

0.72

Activations Density 1.489%

legal and legislative terms or actions

mentions of legislative bills and their specific provisions or effects

No Comments

No Known Activations