INDEX

Explanations

questions ending with a question mark

oai_token-act-pair · gpt-3.5-turbo

questions and inquiries about various topics

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

¥ŀ

-0.84

 eleph

-0.81

 phased

-0.79

 sustained

-0.79

 destro

-0.77

 ende

-0.77

 travers

-0.76

 flowing

-0.75

 flowering

-0.75

 transition

-0.75

POSITIVE LOGITS

Answer

2.47

Well

2.03

Absolutely

1.94

Yes

1.85

Probably

1.83

Wr

1.65

YES

1.65

Sure

1.62

Hmm

1.61

Actually

1.59

Activations Density 0.227%

questions ending with a question mark

questions and inquiries about various topics

No Comments

No Known Activations