INDEX

Explanations

phrases related to political discussions and events

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

accompanied

-0.72

isal

-0.68

foreseen

-0.66

fter

-0.63

agonist

-0.62

cedented

-0.61

Pass

-0.61

aml

-0.60

critical

-0.60

apeshifter

-0.60

POSITIVE LOGITS

 themselves

1.29

selves

0.76

 selves

0.71

 uniforms

0.71

 THEIR

0.71

 fools

0.70

 microphones

0.68

 advertise

0.67

 yourselves

0.67

 confisc

0.67

Activations Density 0.501%

phrases related to political discussions and events

No Comments

No Known Activations