INDEX

Explanations

words related to mysterious or puzzling situations

oai_token-act-pair · gpt-3.5-turbo

references to mystery and ambiguous situations

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

ariat

-0.73

enting

-0.70

ITT

-0.69

 Supported

-0.65

fters

-0.64

 Tend

-0.64

Deploy

-0.64

 externalToEVAOnly

-0.62

ENDED

-0.61

bars

-0.61

POSITIVE LOGITS

 mystery

3.76

 mysteries

2.79

 Mystery

2.50

 Mysteries

2.00

stery

1.73

 mysterious

1.72

 myster

1.71

 puzzle

1.71

 suspense

1.64

 intrigue

1.60

Activations Density 0.022%

words related to mysterious or puzzling situations

references to mystery and ambiguous situations

No Comments

No Known Activations