INDEX

Explanations

phrases related to interacting with content or subscribing to newsletters

oai_token-act-pair · gpt-3.5-turbo

interactive actions or commands related to user engagement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 hurd

-0.82

 vocal

-0.70

 whistle

-0.67

 Catalyst

-0.65

 reckoned

-0.63

 conver

-0.60

 revolving

-0.60

Ĥª

-0.59

 foremost

-0.58

 tant

-0.58

POSITIVE LOGITS

 Cancel

0.76

ename

0.69

 Email

0.67

iful

0.67

Property

0.66

igers

0.66

ById

0.65

iframe

0.62

ribute

0.61

 Thumbnails

0.61

Activations Density 0.027%

phrases related to interacting with content or subscribing to newsletters

interactive actions or commands related to user engagement

No Comments

No Known Activations