INDEX

Explanations

requests for trust in news content

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the reliability of information

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uay

-0.64

 Tsukuyomi

-0.59

luster

-0.58

dan

-0.56

 Marine

-0.56

 Alone

-0.55

 Laurent

-0.55

amaz

-0.54

DonaldTrump

-0.53

 Dunham

-0.53

POSITIVE LOGITS

'll

0.62

ires

0.60

've

0.59

icas

0.57

ikes

0.57

ributes

0.57

iku

0.56

're

0.55

LI

0.54

 weekday

0.53

Activations Density 0.016%

requests for trust in news content

phrases indicating the reliability of information

No Comments

No Known Activations