INDEX

Explanations

references to law enforcement

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

 Tate

-0.73

 Philips

-0.66

Eco

-0.66

Dy

-0.64

 Wast

-0.63

Voy

-0.62

 Churchill

-0.61

 Carbuncle

-0.61

 Seasons

-0.60

 Remastered

-0.60

POSITIVE LOGITS

 enforcement

1.52

suit

1.34

enforcement

1.31

breakers

1.26

 Enforcement

1.25

suits

1.23

lessness

1.18

breaking

1.16

breaker

1.15

 abiding

1.14

Activations Density 0.038%

references to law enforcement

No Comments

No Known Activations