INDEX

Explanations

words or phrases related to quotations or reported speech

oai_token-act-pair · gpt-3.5-turbo

phrases indicating direct quotes or reported speech

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 theirs

-0.56

 creeps

-0.46

 churn

-0.45

 anymore

-0.42

 lame

-0.42

advertising

-0.42

 anywhere

-0.42

).

-0.42

!".

-0.42

.).

-0.42

POSITIVE LOGITS

ccording

0.53

 Miliband

0.51

resa

0.51

 Canaver

0.48

OWER

0.48

isites

0.47

adr

0.45

 Lavrov

0.45

odore

0.45

 Chomsky

0.45

Activations Density 1.802%

words or phrases related to quotations or reported speech

phrases indicating direct quotes or reported speech

No Comments

No Known Activations