INDEX

Explanations

occurrences of unique identifiers or notable phrases

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

conditional statements related to code behavior

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

various contexts and languages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

ьаж

-0.63

Handlung

-0.60

 linkovi

-0.57

GOTREF

-0.57

SequentialGroup

-0.55

BeginContext

-0.53

/**

-0.53

 tajam

-0.52

__":

-0.51

localctx

-0.51

POSITIVE LOGITS

 purpoſe

0.73

 greateſt

0.68

 Jefus

0.65

autoload

0.61

 scattata

0.61

 Monfieur

0.61

 pleaſure

0.61

 RIPRODUZIONE

0.60

 étoit

0.57

ContextHolder

0.56

Activations Density 0.081%

occurrences of unique identifiers or notable phrases

conditional statements related to code behavior

various contexts and languages

No Comments

No Known Activations