INDEX

Explanations

phrases that indicate evidence or examples in discussions or arguments

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

ortium

-0.74

sonian

-0.72

itiz

-0.69

 belonged

-0.68

 intended

-0.67

oided

-0.67

 intending

-0.67

linger

-0.67

 dunno

-0.65

ieties

-0.65

POSITIVE LOGITS

 Figure

0.91

 screenshots

0.86

 Figures

0.84

 graphs

0.83

Fig

0.80

 recent

0.79

 pictures

0.75

 Exhibit

0.75

 numerous

0.73

Fig

0.72

Activations Density 0.097%

phrases that indicate evidence or examples in discussions or arguments

No Comments

No Known Activations