INDEX

Explanations

locations or places

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

Ô

-0.69

agascar

-0.68

"}],"

-0.68

SpaceEngineers

-0.66

=[

-0.65

arella

-0.64

ÙĲ

-0.63

 Advertisement

-0.62

orry

-0.62

Tier

-0.61

POSITIVE LOGITS

 totality

0.79

 acron

0.63

 digits

0.60

 heavens

0.60

liest

0.59

 continents

0.58

 Architects

0.58

 possibilities

0.57

ties

0.57

asses

0.56

Activations Density 0.158%

locations or places

No Comments

No Known Activations