INDEX

Explanations

words related to news broadcasting and media personalities

oai_token-act-pair · gpt-3.5-turbo

references to news hosts and their segments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

byss

-0.83

 ILCS

-0.83

 envis

-0.80

 wards

-0.78

 parcels

-0.74

LTD

-0.73

 labou

-0.73

 signatures

-0.73

ICT

-0.69

 Agric

-0.69

POSITIVE LOGITS

 MSNBC

1.67

 Hannity

1.44

 Colbert

1.42

CNN

1.41

 CNBC

1.40

NBC

1.33

 Kimmel

1.31

NBC

1.31

CNN

1.28

 Comedy

1.26

Activations Density 0.637%

words related to news broadcasting and media personalities

references to news hosts and their segments

No Comments

No Known Activations