INDEX

Explanations

terms related to unique or unusual characteristics

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

pojo | POCO

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

SBATCH

-0.55

ViewInit

-0.50

 consultato

-0.50

MeasureSpec

-0.48

 zagran

-0.47

 grecque

-0.47

 onCancelled

-0.46

makeText

-0.45

 zimowa

-0.45

 administrativa

-0.45

POSITIVE LOGITS

pojo

1.48

Pojo

1.09

Pob

0.59

Poy

0.59

 jsonObject

0.58

Pog

0.57

Po

0.57

 jojo

0.56

oj

0.55

 JSONObject

0.54

Activations Density 0.001%

terms related to unique or unusual characteristics

pojo | POCO

No Comments

No Known Activations