INDEX

Explanations

phrases related to doing work, often of a physical or laborious nature

oai_token-act-pair · gpt-3.5-turbo

phrases related to undesirable or unpleasant tasks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

urated

-0.70

 Sett

-0.69

oor

-0.68

 Bound

-0.68

 Arri

-0.68

 Continued

-0.67

 Courage

-0.67

ãĤ¦ãĤ¹

-0.67

 Returning

-0.66

eming

-0.66

POSITIVE LOGITS

 grunt

0.89

 chores

0.83

 homework

0.80

strip

0.75

 differently

0.74

 offline

0.74

 rehab

0.71

 migrate

0.68

 thing

0.68

 experiment

0.68

Activations Density 0.256%

phrases related to doing work, often of a physical or laborious nature

phrases related to undesirable or unpleasant tasks

No Comments

No Known Activations