INDEX

Explanations

tweets and references to Twitter in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ageing

-0.69

 substitutes

-0.65

 delinquent

-0.63

 cooperative

-0.63

tein

-0.62

 totality

-0.62

 scores

-0.62

 aging

-0.61

quartered

-0.61

 defense

-0.60

POSITIVE LOGITS

twitter

1.14

://

0.89

cdn

0.83

#$

0.81

youtu

0.81

ðŁ

0.80

 Appears

0.77

âĢİ

0.75

youtube

0.74

URL

0.73

Activations Density 0.010%

tweets and references to Twitter in the text

No Comments

No Known Activations