INDEX

Explanations

technical documentation, code snippets and their compilation, and anything about hardware product specifications

oai_token-act-pair · gemini-2.0-flash

actions or processes

np_max-act-logits · gemini-2.0-flash

say "prepositions" Method used: 2 — because the tokens after the max-activating tokens are primarily prepositions/particles; say "prepositions"

np_max-act-logits · gpt-5-mini Triggered by @chenshw0109

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Rüyada

-0.63

скажите

-0.56

页面存档备份

-0.56

rungsseite

-0.56

BibitemShut

-0.53

tifies

-0.50

发表于

-0.49

 générations

-0.49

 дописавши

-0.47

 disambiguazione

-0.47

POSITIVE LOGITS

 displays

1.33

 runs

1.31

 counts

1.30

 feeds

1.29

 acts

1.28

 stands

1.27

 supports

1.23

 spans

1.23

 uses

1.22

 flows

1.20

Activations Density 37.366%

technical documentation, code snippets and their compilation, and anything about hardware product specifications

actions or processes

say "prepositions" Method used: 2 — because the tokens after the max-activating tokens are primarily prepositions/particles; say "prepositions"

No Comments

No Known Activations