INDEX

Explanations

phrases that indicate quotes or dialogue

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 conj

-0.71

 Vaugh

-0.68

 listed

-0.66

azo

-0.65

 Pavel

-0.63

 specified

-0.63

 scheduled

-0.62

 paired

-0.61

 configured

-0.61

 mund

-0.61

POSITIVE LOGITS

Cause

1.12

Very

1.04

Absolutely

1.02

Extreme

0.98

Strong

0.95

Political

0.92

Really

0.90

cause

0.90

Going

0.89

Better

0.88

Activations Density 0.015%

phrases that indicate quotes or dialogue

No Comments

No Known Activations