INDEX

Explanations

informative terms and prompts, likely used in website articles or blog posts

oai_token-act-pair · gpt-3.5-turbo

the presence of non-standard tokens or placeholders in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

stood

-0.72

#$

-0.66

ppard

-0.63

forth

-0.61

 spont

-0.60

 Borders

-0.59

manship

-0.58

minist

-0.58

 Overse

-0.57

AFTA

-0.56

POSITIVE LOGITS

actionDate

0.81

0.80

][

0.80

 TEXT

0.77

 guiActiveUnfocused

0.76

"]=>

0.72

quickShipAvailable

0.70

];

0.69

 async

0.68

 guiActive

0.67

Activations Density 0.187%

informative terms and prompts, likely used in website articles or blog posts

the presence of non-standard tokens or placeholders in the text

No Comments

No Known Activations