INDEX

Explanations

full names and organizations within a context of communication and discussion

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

SHARE

-0.64

 testament

-0.64

SIZE

-0.63

 dope

-0.60

 appropriation

-0.59

xp

-0.56

CHECK

-0.55

 signature

-0.54

 initials

-0.54

 mole

-0.53

POSITIVE LOGITS

pai

0.82

arre

0.80

quart

0.79

rador

0.76

christ

0.75

aghd

0.72

illard

0.70

instein

0.70

terson

0.70

gomery

0.69

Activations Density 0.091%

full names and organizations within a context of communication and discussion

No Comments

No Known Activations