INDEX

Explanations

conditional phrases and assessments

oai_token-act-pair · gpt-4o-mini Triggered by @bot

Followed by "full review" or "full" (of something)

np_acts-logits-general · gemini-2.0-flash

full review

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kháu

-0.52

webElementXpaths

-0.49

 Picchu

-0.43

 artikkelen

-0.43

 мәкал

-0.42

 defaultstate

-0.42

hands

-0.42

 Offisielt

-0.41

makeConstraints

-0.41

 HANDS

-0.41

POSITIVE LOGITS

too

0.53

これも

0.52

 again

0.51

 Again

0.46

Again

0.42

 ebenfalls

0.41

again

0.41

 дописавши

0.41

 yine

0.40

こちらも

0.40

Activations Density 1.190%

conditional phrases and assessments

Followed by "full review" or "full" (of something)

full review

No Comments

No Known Activations