INDEX

Explanations

numerical values

oai_token-act-pair · gpt-3.5-turbo

terms indicating frequency and inquiry, especially regarding additional information

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

ãĥ©ãĥ³

-0.59

vernment

-0.57

rals

-0.56

 Palest

-0.56

 glac

-0.54

etus

-0.54

 Guant

-0.54

ccording

-0.54

 purported

-0.53

 boosters

-0.52

POSITIVE LOGITS

endif

0.83

ONSORED

0.73

--------------------------------------------------------

0.62

 Also

0.62

 NOTE

0.61

 Delete

0.61

 Related

0.61

 furthermore

0.59

 PHOTO

0.58

else

0.58

Activations Density 0.184%

numerical values

terms indicating frequency and inquiry, especially regarding additional information

No Comments

No Known Activations