INDEX

Explanations

instances of negations and instructions

oai_token-act-pair · gpt-3.5-turbo

negations and conditions regarding actions or permissions

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

pires

-0.82

 Born

-0.74

BuyableInstoreAndOnline

-0.71

quartered

-0.68

dominated

-0.68

nutrition

-0.67

 millenn

-0.67

ãĤ¼ãĤ¦ãĤ¹

-0.67

htaking

-0.66

 Moroc

-0.65

POSITIVE LOGITS

 delete

1.33

 modify

1.28

 refund

1.27

 revoke

1.26

 retrieve

1.25

 reuse

1.25

 remove

1.24

 amend

1.24

 recomp

1.23

 proceed

1.21

Activations Density 0.330%

instances of negations and instructions

negations and conditions regarding actions or permissions

No Comments

No Known Activations