INDEX

Explanations

phrases related to software and technology

oai_token-act-pair · gpt-3.5-turbo

instances of statistical or numerical data results

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 glim

-0.83

 quir

-0.77

aleigh

-0.76

flo

-0.75

 rodent

-0.75

 swat

-0.74

 babys

-0.74

 bourbon

-0.74

 shorthand

-0.74

 trespass

-0.73

POSITIVE LOGITS

 Therefore

1.82

 Hence

1.75

 Moreover

1.73

 Besides

1.69

Therefore

1.56

 Furthermore

1.55

 Afterwards

1.54

 Secondly

1.49

 Regarding

1.49

 Anyway

1.48

Activations Density 0.626%

phrases related to software and technology

instances of statistical or numerical data results

No Comments

No Known Activations