INDEX

Explanations

news agency abbreviations

oai_token-act-pair · gpt-3.5-turbo

references to news organizations or agencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Crow

-0.76

 Mill

-0.71

 closed

-0.69

 Twin

-0.65

 enclosed

-0.65

 Doll

-0.65

 splitting

-0.64

 rooms

-0.64

Bot

-0.63

 Twins

-0.62

POSITIVE LOGITS

AFP

4.19

AFP

2.09

Reuters

1.90

EPA

1.64

Getty

1.44

Bloomberg

1.40

CNN

1.34

Flickr

1.33

AP

1.20

NPR

1.19

Activations Density 0.010%

news agency abbreviations

references to news organizations or agencies

No Comments

No Known Activations