INDEX

Explanations

adverbs indicating time or frequency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

matic

-0.71

enary

-0.66

 genuinely

-0.65

osp

-0.64

unn

-0.63

othal

-0.63

ente

-0.62

bern

-0.62

ums

-0.62

Really

-0.62

POSITIVE LOGITS

 evidenced

0.81

 Micha

0.69

aukee

0.68

 Chap

0.68

 eyeing

0.66

 Cyrus

0.66

 served

0.66

 manifested

0.64

 resides

0.62

 Pengu

0.61

Activations Density 0.020%

adverbs indicating time or frequency

No Comments

No Known Activations