INDEX

Explanations

terms associated with intelligence and cognitive ability

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

dule

-0.81

soDeliveryDate

-0.77

 Soda

-0.74

raine

-0.70

atern

-0.69

risome

-0.67

pee

-0.66

kef

-0.65

zl

-0.65

ibur

-0.65

POSITIVE LOGITS

 quot

1.23

elligence

0.90

ibility

0.77

 enhancement

0.72

 retention

0.71

 deficiency

0.70

smanship

0.70

IQ

0.69

agents

0.68

 superiority

0.67

Activations Density 0.014%

terms associated with intelligence and cognitive ability

No Comments

No Known Activations