INDEX

Explanations

references to governmental and legal processes

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

Punctuation followed by discourse markers/conjunctions

np_acts-logits-general · gemini-2.0-flash

prohibitions and conditions

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_24/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

Clik

-0.56

 Occasionally

-0.56

Occasionally

-0.55

 Toujours

-0.53

 toujours

-0.53

 ModelExpression

-0.53

 occasionally

-0.52

Toujours

-0.51

sempre

-0.50

WindowListener

-0.49

POSITIVE LOGITS

 Allowing

0.85

 banning

0.82

if

0.79

 unless

0.78

 Unless

0.77

If

0.77

 depriving

0.76

 abol

0.76

 removing

0.74

このまま

0.74

Activations Density 0.357%

references to governmental and legal processes

Punctuation followed by discourse markers/conjunctions

prohibitions and conditions

No Known Activations