INDEX

Explanations

references to the duration of timeouts or alerts in programming contexts

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

references to toast notifications in programming code

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

short and long

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

MessageTagHelper

-0.69

NameInMap

-0.64

estacks

-0.60

ItemBackground

-0.56

saraba

-0.56

 GeoNames

-0.56

erca

-0.56

rungsseite

-0.55

SourceChecksum

-0.55

rzej

-0.53

POSITIVE LOGITS

SHORT

1.04

 short

0.85

short

0.78

 SHORT

0.76

LONG

0.76

 Short

0.69

Short

0.68

 long

0.59

 shor

0.59

extAlignment

0.59

Activations Density 0.003%

references to the duration of timeouts or alerts in programming contexts

references to toast notifications in programming code

short and long

No Comments

No Known Activations