INDEX

Explanations

the word "that" in various contexts and usages

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

 Corps

-0.62

 Riding

-0.62

Uk

-0.60

 guessing

-0.60

 Vend

-0.55

 Afgh

-0.54

Ide

-0.53

 Guess

-0.53

Soc

-0.53

Tax

-0.52

POSITIVE LOGITS

 resulted

1.03

 culminated

0.97

 accompanies

0.96

 spans

0.95

 preceded

0.95

 exceeds

0.94

 includes

0.93

 lasted

0.92

 span

0.88

 arose

0.87

Activations Density 0.187%

the word "that" in various contexts and usages

No Comments

No Known Activations