INDEX

Explanations

phrases related to politics and government status

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

tein

-0.69

fell

-0.68

 plagiar

-0.67

 Sisters

-0.62

HQ

-0.62

 Xuan

-0.61

ortment

-0.61

 predecessor

-0.60

 Sabb

-0.59

 Mansion

-0.59

POSITIVE LOGITS

 clam

1.15

 overwhelmingly

0.93

 dissatisfied

0.93

 perce

0.92

 hungry

0.89

 perceive

0.85

 revolt

0.84

 disillusion

0.83

 accustomed

0.82

 electing

0.81

Activations Density 0.679%

phrases related to politics and government status

No Comments

No Known Activations