INDEX

Explanations

informational content related to police investigations, appeals for information, and rewards for details

oai_token-act-pair · gpt-3.5-turbo

phrases indicating requests for information or contact details in a law enforcement context

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

 monopol

-0.70

BILL

-0.70

å¥

-0.70

obook

-0.69

-0.67

Pin

-0.67

-0.64

utterstock

-0.64

flo

-0.63

POSITIVE LOGITS

 contacted

0.81

 confir

0.75

 Investigators

0.74

 enqu

0.73

 spokesman

0.71

 condolences

0.69

 corrections

0.69

details

0.69

 emailed

0.68

 investigator

0.68

Activations Density 1.251%

informational content related to police investigations, appeals for information, and rewards for details

phrases indicating requests for information or contact details in a law enforcement context

No Comments

No Known Activations