INDEX

Explanations

terms related to strong commitments or dedications

oai_token-act-pair · gpt-3.5-turbo

expressions of commitment

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

adish

-0.70

 complexion

-0.69

quer

-0.67

avis

-0.67

atin

-0.65

ebin

-0.64

ocks

-0.64

ourning

-0.63

ox

-0.63

phe

-0.62

POSITIVE LOGITS

 commitment

1.03

 allegiance

0.91

 commitments

0.85

irmation

0.83

gence

0.81

venant

0.74

 obligation

0.73

duty

0.71

ment

0.68

 Commit

0.67

Activations Density 0.018%

terms related to strong commitments or dedications

expressions of commitment

No Comments

No Known Activations