INDEX

Explanations

news reports or articles mentioning specific events or information

oai_token-act-pair · gpt-3.5-turbo

phrases mentioning "reports."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 wedge

-0.73

 palate

-0.66

egal

-0.62

 ILCS

-0.61

east

-0.60

animate

-0.60

oan

-0.60

aughs

-0.60

å§«

-0.60

 Klux

-0.59

POSITIVE LOGITS

 reports

0.88

books

0.83

etter

0.79

udo

0.75

Reports

0.72

Reporting

0.72

reports

0.72

 briefings

0.71

 Reports

0.70

uggest

0.70

Activations Density 0.029%

news reports or articles mentioning specific events or information

phrases mentioning "reports."

No Comments

No Known Activations