INDEX

Explanations

instances of location-related references or spatial indicators within the text

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

location marker "here"

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

The neuron mainly detects deictic/location or discourse markers like "here" and "there" (words that refer to the current context or location in the text).

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

ben

-0.29

setViewName

-0.29

 בז

-0.28

ilaire

-0.28

oks

-0.28

orianCalendar

-0.28

multirow

-0.27

GeneratedCode

-0.27

sta

-0.26

 Conservancy

-0.26

POSITIVE LOGITS

 ujednoznacz

0.69

tanleria

0.63

Jeografia

0.62

 بيها

0.60

Искәрмәләр

0.60

IsMutable

0.59

+#+

0.58

 dezelve

0.56

 المعيارى

0.56

 spesa

0.55

Activations Density 1.649%

instances of location-related references or spatial indicators within the text

location marker "here"

The neuron mainly detects deictic/location or discourse markers like "here" and "there" (words that refer to the current context or location in the text).

No Comments

No Known Activations

instances of location-related references or spatial indicators within the text

location marker "here"

The neuron mainly detects deictic/location or discourse markers like "here" and "there" (words that refer to the current context or location in the text).

No Comments

No Known Activations