INDEX

Explanations

words related to legal or criminal activities from court records

oai_token-act-pair · gpt-3.5-turbo

references to legal documentation and court records

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

ãĥĥãĤ¯

-0.82

atorial

-0.82

ãĤª

-0.74

ãĤ§

-0.74

otin

-0.73

otal

-0.73

atism

-0.72

obyl

-0.70

Grad

-0.70

cos

-0.69

POSITIVE LOGITS

 affidavit

0.91

 affidav

0.75

Sgt

0.71

KEY

0.70

 troopers

0.68

idav

0.67

 filings

0.67

 Dempsey

0.66

 records

0.66

 backups

0.65

Activations Density 0.209%

words related to legal or criminal activities from court records

references to legal documentation and court records

No Comments

No Known Activations