INDEX

Explanations

details related to news analysis and feedback

oai_token-act-pair · gpt-3.5-turbo

structured information, particularly in the context of analysis and feedback in academic or formal writing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 disadvant

-0.67

 destro

-0.66

 subsistence

-0.55

bom

-0.53

 Brah

-0.50

 undermin

-0.49

 pract

-0.48

chuk

-0.47

 helicop

-0.47

 reckoned

-0.46

POSITIVE LOGITS

isode

0.57

 âĢº

0.57

FAQ

0.53

 blog

0.52

 Transcript

0.50

º

0.48

ython

0.48

 Editorial

0.48

 podcast

0.48

 reader

0.47

Activations Density 2.022%

details related to news analysis and feedback

structured information, particularly in the context of analysis and feedback in academic or formal writing

No Comments

No Known Activations