INDEX

Explanations

legal terms and court-related decisions

oai_token-act-pair · gpt-3.5-turbo

terms related to court rulings and legal decisions

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

eworld

-0.85

ohyd

-0.74

ongh

-0.74

ople

-0.71

 goose

-0.71

dden

-0.68

 airst

-0.68

pload

-0.67

udeau

-0.66

 labou

-0.65

POSITIVE LOGITS

making

0.88

Sup

0.88

 rulings

0.83

 upholding

0.82

Jud

0.82

Judge

0.80

sup

0.79

 legalizing

0.79

 Cosponsors

0.78

maker

0.75

Activations Density 0.059%

legal terms and court-related decisions

terms related to court rulings and legal decisions

No Comments

No Known Activations