INDEX

Explanations

instances where actions or decisions are being postponed or delayed

oai_token-act-pair · gpt-3.5-turbo

terms related to the concept of deferral or postponement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Passion

-0.75

opher

-0.73

ARS

-0.72

ovych

-0.72

 mole

-0.72

++++++++

-0.71

ī

-0.70

bern

-0.68

å·

-0.68

Īè

-0.67

POSITIVE LOGITS

 deferred

1.46

 defer

1.34

erred

1.26

 adjud

0.84

 warr

0.83

minist

0.80

ptive

0.77

 gratification

0.77

 assignments

0.75

 acknow

0.73

Activations Density 0.008%

instances where actions or decisions are being postponed or delayed

terms related to the concept of deferral or postponement

No Comments

No Known Activations