INDEX

Explanations

references to political affiliations, particularly "R" for Republican and "D" for Democrat

oai_token-act-pair · gpt-4o-mini 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

 supplies

-0.67

 disclaim

-0.65

 compilation

-0.64

 gratification

-0.63

 Mandela

-0.63

 Guinness

-0.63

 transports

-0.62

 variables

-0.62

 exploits

-0.62

 tutorials

-0.61

POSITIVE LOGITS

ICH

0.97

appa

0.93

ollo

0.93

aska

0.92

)'

0.91

iffe

0.90

NJ

0.90

UST

0.86

ump

0.85

achus

0.85

Activations Density 0.018%

references to political affiliations, particularly "R" for Republican and "D" for Democrat

No Comments

No Known Activations