INDEX

Explanations

symbols and comments used in programming or markup languages

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

API operation verbs

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

fromnode

-0.59

 betweenstory

-0.54

 SwitchCompat

-0.53

ThroughAttribute

-0.51

 <<<<<<<<<<<<<<

-0.51

principalTable

-0.49

 Ahnung

-0.49

giaan

-0.49

Билгалдахарш

-0.47

ołnier

-0.47

POSITIVE LOGITS

<bos>

0.45

 JADX

0.44

httphttps

0.42

Portale

0.42

Становништво

0.40

<th>

0.39

Matters

0.35

 surla

0.35

 makeStyles

0.34

 Matters

0.34

Activations Density 0.001%

symbols and comments used in programming or markup languages

API operation verbs

No Comments

No Known Activations