INDEX

Explanations

phrases involving specific interactions with characters or specific activities

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

ovich

-0.74

 ende

-0.72

omsky

-0.72

etary

-0.69

 conclud

-0.69

.–

-0.67

pring

-0.67

fair

-0.66

venant

-0.66

bia

-0.65

POSITIVE LOGITS

 friends

1.16

stood

1.16

 coworkers

1.07

 pals

1.05

 colleagues

1.03

 strangers

0.98

 classmates

0.97

 buddies

0.96

 fellow

0.96

 teammates

0.95

Activations Density 0.190%

phrases involving specific interactions with characters or specific activities

No Comments

No Known Activations