INDEX

Explanations

phrases related to taking care of or managing something

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

coni

-0.62

 Chili

-0.59

Kings

-0.58

chell

-0.55

 Sands

-0.55

activation

-0.55

eru

-0.54

chrom

-0.54

eele

-0.53

 rage

-0.53

POSITIVE LOGITS

taker

1.14

giving

0.88

tesy

0.82

 largeDownload

0.76

fully

0.75

ktop

0.75

tes

0.74

taking

0.73

 âĶľ

0.73

worn

0.72

Activations Density 0.012%

phrases related to taking care of or managing something

No Comments

No Known Activations