INDEX

Explanations

emotional and personal reflections or expressions

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

mony

-0.66

noon

-0.65

mon

-0.62

ierrez

-0.60

hma

-0.59

artisan

-0.57

jury

-0.57

geoning

-0.57

ikers

-0.56

rontal

-0.56

POSITIVE LOGITS

 fuss

0.95

 entails

0.82

 wrought

0.81

 boils

0.78

 entail

0.78

 Means

0.77

 Learned

0.77

 meant

0.77

DERR

0.76

ãĤ´

0.75

Activations Density 0.741%

emotional and personal reflections or expressions

No Comments

No Known Activations