INDEX

Explanations

content related to news alerts and reports

oai_token-act-pair · gpt-3.5-turbo

phrases related to news updates and reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

iple

-0.65

 coerc

-0.64

 appell

-0.62

awa

-0.60

igl

-0.60

âĦ¢:

-0.57

 hardness

-0.56

IPM

-0.56

 somew

-0.55

amiya

-0.55

POSITIVE LOGITS

 mornings

0.70

 Breaking

0.67

cean

0.64

stories

0.61

 Watergate

0.61

 stories

0.60

vernight

0.58

 weekday

0.58

 alerts

0.57

 Kavanaugh

0.57

Activations Density 0.044%

content related to news alerts and reports

phrases related to news updates and reports

No Comments

No Known Activations