INDEX

Explanations

instances of numerical references or high-stakes figures related to achievement or recognition

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

section titles or continuations

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

 AssemblyCulture

-0.68

+#+#

-0.56

 تانيه

-0.55

 بيها

-0.50

Rujuakan

-0.50

Personensuche

-0.49

 borbo

-0.49

]));

-0.49

unknownFields

-0.48

"]))

-0.47

POSITIVE LOGITS

ervlak

0.40

новништво

0.39

 AssemblyProduct

0.35

 xúc

0.35

 препратки

0.34

 handleMessage

0.32

StringCopy

0.31

AttributeSet

0.31

0.30

setContentType

0.30

Activations Density 0.072%

instances of numerical references or high-stakes figures related to achievement or recognition

section titles or continuations

No Comments

No Known Activations