INDEX

Explanations

specific references to news agencies, specifically "AP" (Associated Press)

oai_token-act-pair · gpt-3.5-turbo

references to the Associated Press (AP) news agency

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

worms

-0.74

 hither

-0.69

 knit

-0.68

worm

-0.68

 ninety

-0.66

 Lyme

-0.66

 stool

-0.60

ankind

-0.60

 inval

-0.59

 dressing

-0.59

POSITIVE LOGITS

AP

3.72

AP

2.22

 Associated

1.61

AFP

1.52

APS

1.47

SAP

1.42

EPA

1.39

 Reuters

1.39

APR

1.38

AAP

1.37

Activations Density 0.016%

specific references to news agencies, specifically "AP" (Associated Press)

references to the Associated Press (AP) news agency

No Comments

No Known Activations