INDEX

Explanations

instances of time indicators

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.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

BIP

-0.70

pel

-0.65

 Sins

-0.63

acus

-0.61

puter

-0.61

initely

-0.60

ichick

-0.59

CHQ

-0.59

mic

-0.57

artisan

-0.57

POSITIVE LOGITS

 thereafter

1.29

aneously

1.12

 afterwards

0.95

 afterward

0.94

 enough

0.92

 followed

0.79

 forgotten

0.79

ened

0.77

est

0.76

rise

0.74

Activations Density 0.547%

instances of time indicators

No Comments

No Known Activations