INDEX

Explanations

mentions of specific places or locations, especially related to conflict or military actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 crochet

-0.71

 Roller

-0.66

 erection

-0.65

 Entered

-0.65

=-=-=-=-

-0.64

xp

-0.63

 gears

-0.62

acebook

-0.62

livious

-0.60

EVE

-0.60

POSITIVE LOGITS

azeera

1.14

aghd

1.08

awi

1.01

qqa

0.95

aq

0.92

abi

0.91

ouf

0.89

aeda

0.88

andi

0.87

aba

0.85

Activations Density 0.035%

mentions of specific places or locations, especially related to conflict or military actions

No Comments

No Known Activations