INDEX

Explanations

phrases related to behind the scenes and discussions happening behind the scenes

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

RIC

-0.76

arily

-0.67

 anew

-0.65

tm

-0.62

fy

-0.60

ACY

-0.59

nces

-0.59

ERG

-0.59

nu

-0.59

fully

-0.58

POSITIVE LOGITS

 scenes

1.18

 curtain

1.11

 Scenes

0.95

 curtains

0.89

 veil

0.87

 wheel

0.83

 curve

0.81

 desk

0.78

scenes

0.78

 lines

0.74

Activations Density 0.180%

phrases related to behind the scenes and discussions happening behind the scenes

No Comments

No Known Activations