INDEX

Explanations

information related to emails, particularly their content and recipients

oai_token-act-pair · gpt-3.5-turbo

references to leaked emails and communications

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

iol

-0.76

udic

-0.76

 Trials

-0.68

PI

-0.66

 Coverage

-0.66

 Flavoring

-0.66

iculture

-0.64

ivia

-0.63

ovy

-0.63

enment

-0.63

POSITIVE LOGITS

 messages

0.99

 Messages

0.89

 dated

0.88

 overheard

0.88

 messaging

0.87

 relayed

0.83

 handwritten

0.81

 communications

0.81

dated

0.80

 circulated

0.80

Activations Density 0.289%

information related to emails, particularly their content and recipients

references to leaked emails and communications

No Comments

No Known Activations