INDEX

Explanations

statistics or numerical values embedded in a text

oai_token-act-pair · gpt-3.5-turbo

references to statistical measurements or margins of error

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

and

-0.56

etc

-0.53

 Houses

-0.52

 portfolios

-0.52

or

-0.49

 fleets

-0.49

 vans

-0.49

 respectively

-0.49

 museums

-0.49

 Machines

-0.47

POSITIVE LOGITS

«

0.65

»

0.64

uch

0.63

chest

0.61

fits

0.60

percent

0.60

onis

0.59

ãĤ§

0.59

ps

0.59

³

0.58

Activations Density 0.752%

statistics or numerical values embedded in a text

references to statistical measurements or margins of error

No Comments

No Known Activations