INDEX

Explanations

patterns involving numerical data or statistics

oai_token-act-pair · gpt-4o-mini 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 trance

-0.61

 compulsion

-0.59

 alignment

-0.58

 trail

-0.58

ĪĴ

-0.58

 reflex

-0.58

 loophole

-0.56

 ratification

-0.55

 tele

-0.54

rat

-0.53

POSITIVE LOGITS

ppe

0.80

lycer

0.76

ÏĤ

0.67

oteric

0.66

 Thumbnails

0.66

AppData

0.65

vati

0.64

}}}

0.64

emetery

0.64

oppers

0.64

Activations Density 0.047%

patterns involving numerical data or statistics

No Comments

No Known Activations