INDEX

Explanations

titles or names of creative works, particularly in popular culture

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

 detectors

-0.84

 corridors

-0.80

 scanners

-0.80

 levers

-0.78

lees

-0.78

 belts

-0.77

 formations

-0.77

 brackets

-0.76

 establishments

-0.76

 faculties

-0.75

POSITIVE LOGITS

 Called

1.20

 Named

1.00

 Perspective

0.98

 Reborn

0.97

 Tale

0.96

 Worth

0.96

 Difference

0.93

illion

0.88

 Worse

0.87

ccording

0.83

Activations Density 0.102%

titles or names of creative works, particularly in popular culture

No Comments

No Known Activations