INDEX

Explanations

punctuation marks, specifically periods

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

enum types and attributes

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

 Offisielt

-0.66

 Inscrivez

-0.59

 desmotivaciones

-0.59

guiente

-0.55

 auroit

-0.54

 sportowe

-0.53

 logistique

-0.51

 supplémentaires

-0.50

 ainfi

-0.50

 sánchez

-0.50

POSITIVE LOGITS

0.66

(&:

0.65

RTDA

0.64

(_.

0.61

/\.

0.60

`.

0.58

[.

0.57

(.

0.57

~/.

0.57

 &___

0.57

Activations Density 0.006%

punctuation marks, specifically periods

enum types and attributes

No Comments

No Known Activations