INDEX

Explanations

mentions of levels of a certain attribute or characteristic

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

vous

-0.69

 Andromeda

-0.69

orial

-0.69

fect

-0.68

MRI

-0.67

sov

-0.67

enegger

-0.66

ILLE

-0.66

ãĥĹ

-0.66

oute

-0.66

POSITIVE LOGITS

 crossings

0.82

xual

0.77

 assurance

0.73

 levels

0.71

chart

0.70

level

0.69

 attained

0.67

 attainment

0.66

 awareness

0.65

 rise

0.64

Activations Density 0.081%

mentions of levels of a certain attribute or characteristic

No Comments

No Known Activations