INDEX

Explanations

instances of defined numerical values or object identifiers in structured data

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

programming keywords and identifiers

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

niſſe

-0.90

iſchen

-0.84

ſſung

-0.83

⟬

-0.80

<unused8>

-0.80

<unused42>

-0.79

<unused43>

-0.79

<unused41>

-0.79

[@BOS@]

-0.79

<unused14>

-0.79

POSITIVE LOGITS

↵↵

0.35

0.34

0.33

 theo

0.32

 previous

0.29

 Tempel

0.28

0.27

 heer

0.27

Activations Density 0.011%

instances of defined numerical values or object identifiers in structured data

programming keywords and identifiers

No Comments

No Known Activations