INDEX

Explanations

words containing the substring "neg" often in the context of criticism or negativity

oai_token-act-pair · gpt-3.5-turbo

negative connotations or expressions related to negativity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

BuyableInstoreAndOnline

-0.91

=-=-=-=-=-=-=-=-

-0.87

ORGE

-0.85

DragonMagazine

-0.81

NetMessage

-0.80

OHN

-0.79

realDonaldTrump

-0.77

Reviewer

-0.74

Assembly

-0.72

 Metatron

-0.72

POSITIVE LOGITS

otiation

1.41

oti

1.30

neg

1.17

atives

1.06

rito

0.93

asonic

0.85

lect

0.84

ativity

0.82

lected

0.81

rals

0.80

Activations Density 0.009%

words containing the substring "neg" often in the context of criticism or negativity

negative connotations or expressions related to negativity

No Comments

No Known Activations