INDEX

Explanations

specific phrases or terms related to content updates and features on websites and applications

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

common phrases

np_acts-logits-general · gemini-2.5-flash-lite

the neuron responds to content-bearing or topical words (important nouns, verbs, pronouns and discourse markers) rather than function or filler tokens.

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

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

 ProtoMessage

-0.71

setVerticalGroup

-0.67

 ویکی‌پدی

-0.65

 հղումներ

-0.61

oneofs

-0.59

 Савезне

-0.59

 tiguan

-0.58

Personendaten

-0.58

WebElementEntity

-0.58

 saites

-0.58

POSITIVE LOGITS

 Masyarakat

0.29

 nahilalakip

0.28

 eletrônico

0.27

 Erscheinung

0.26

 محفوظة

0.26

 akik

0.26

 jezelf

0.26

 Year

0.25

 jspb

0.25

 usage

0.25

Activations Density 0.160%

specific phrases or terms related to content updates and features on websites and applications

common phrases

the neuron responds to content-bearing or topical words (important nouns, verbs, pronouns and discourse markers) rather than function or filler tokens.

No Comments

No Known Activations

specific phrases or terms related to content updates and features on websites and applications

common phrases

the neuron responds to content-bearing or topical words (important nouns, verbs, pronouns and discourse markers) rather than function or filler tokens.

No Comments

No Known Activations