INDEX

Explanations

descriptions or discussions on processes rather than end results

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the process over the product in various contexts

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

)."

-0.64

)"

-0.64

 please

-0.64

etc

-0.61

?).

-0.61

tho

-0.61

LOL

-0.60

;}

-0.60

}"

-0.59

lol

-0.58

POSITIVE LOGITS

 geographically

0.76

 heterogeneity

0.75

 geography

0.68

 fundamentally

0.67

etheless

0.64

 extraordinarily

0.63

 pree

0.62

 itself

0.61

 disag

0.61

 generational

0.61

Activations Density 1.171%

descriptions or discussions on processes rather than end results

phrases emphasizing the process over the product in various contexts

No Comments

No Known Activations