INDEX

Explanations

email related phrases and notifications, specifically mentioning news and promotions

oai_token-act-pair · gpt-3.5-turbo

references to news-related content and updates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ovember

-0.62

appa

-0.61

wagen

-0.57

 neigh

-0.53

 Logged

-0.52

 Devils

-0.51

etts

-0.51

 denomin

-0.50

Peg

-0.50

 wors

-0.50

POSITIVE LOGITS

month

0.64

bytes

0.62

breaking

0.59

 content

0.58

grain

0.56

Trend

0.56

email

0.55

relevant

0.54

codes

0.54

 Kavanaugh

0.53

Activations Density 0.014%

email related phrases and notifications, specifically mentioning news and promotions

references to news-related content and updates

No Comments

No Known Activations