INDEX

Explanations

terms related to data management and processing in a statistical or computational context

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

language and numerical concepts

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

 […]

-0.56

This

-0.46

‘

-0.44

-0.43

It

-0.42

‘

-0.39

â

-0.39

â

-0.39

In

-0.39

The

-0.39

POSITIVE LOGITS

Portály

0.77

دانشنامهٔ

0.71

 betweenstory

0.67

تقاوى

0.65

Portale

0.63

 محفوظة

0.60

 témoig

0.59

 للاسماء

0.59

 tartalomajánló

0.59

ConstraintMaker

0.58

Activations Density 0.051%

terms related to data management and processing in a statistical or computational context

language and numerical concepts

No Comments

No Known Activations