INDEX

Explanations

mentions of the word "permanent."

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

edin

-0.96

gerald

-0.89

ean

-0.87

ramid

-0.80

rahim

-0.77

aucuses

-0.76

maxwell

-0.76

DragonMagazine

-0.76

ymes

-0.76

heit

-0.75

POSITIVE LOGITS

 fixture

0.92

 residence

0.92

manent

0.85

ITY

0.85

 permanent

0.84

 residency

0.81

 injunction

0.81

 resident

0.81

 marker

0.79

 tenancy

0.76

Activations Density 0.019%

mentions of the word "permanent."

No Comments

No Known Activations