INDEX

Explanations

code snippets and programming syntax elements

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

code placeholders and references

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 snippetHide

-0.55

IVEREF

-0.54

ValueGenerated

-0.49

ChildScrollView

-0.48

 Italijani

-0.47

@"/

-0.47

 BoxDecoration

-0.45

OGND

-0.45

RTCF

-0.45

 ſind

-0.44

POSITIVE LOGITS

 resourceCulture

0.42

isation

0.40

 Laurel

0.38

 substantive

0.36

ization

0.36

GeoNames

0.36

ve

0.35

kenalkan

0.35

 Sexton

0.35

 useSelector

0.34

Activations Density 0.069%

code snippets and programming syntax elements

code placeholders and references

No Comments

No Known Activations