INDEX

Explanations

equipment-related terms, such as hardware and instructions

oai_token-act-pair · gpt-3.5-turbo

references to technological hardware and related discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 âĢ

-0.83

âĢ

-0.81

Ö

-0.70

Ö¼

-0.69

SPONSORED

-0.69

à©

-0.65

ÂŃ

-0.63

âĢİ

-0.62

 à¨

-0.62

ãĥŁ

-0.61

POSITIVE LOGITS

1.63

!".

1.63

!.

1.62

;)

1.53

:)

1.49

!).

1.49

:-)

1.45

!'

1.45

!!!

1.43

!!

1.43

Activations Density 1.000%

equipment-related terms, such as hardware and instructions

references to technological hardware and related discussions

No Comments

No Known Activations