INDEX

Explanations

code structure and programmatic elements

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

types of formats

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

Personendaten

-1.00

parsedMessage

-0.85

ConstraintMaker

-0.84

httphttps

-0.81

ViewFeatures

-0.80

rrggbb

-0.80

exitRule

-0.77

MLLoader

-0.74

MigrationBuilder

-0.74

 CreateTagHelper

-0.73

POSITIVE LOGITS

 Eventually

0.36

 Remember

0.36

 known

0.35

 dostępne

0.34

 tanke

0.33

 available

0.33

 They

0.33

Eventually

0.32

 Having

0.32

 запом

0.32

Activations Density 0.021%

code structure and programmatic elements

types of formats

No Comments

No Known Activations