INDEX

Explanations

technical instructions or guides

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

gom

-0.68

bies

-0.66

Footnote

-0.65

lie

-0.65

selves

-0.64

ievers

-0.62

aneers

-0.62

opez

-0.61

chev

-0.61

 automakers

-0.60

POSITIVE LOGITS

NOT

0.78

 definitely

0.77

 unofficial

0.75

 informative

0.71

official

0.70

 VERY

0.70

 Scythe

0.69

 Mahjong

0.68

 easiest

0.67

 amazing

0.66

Activations Density 0.083%

technical instructions or guides

No Comments

No Known Activations