INDEX

Explanations

utter

np_max-act · gemini-2.0-flash

The neuron activates on occurrences of the word “query” (and closely related tokens like “user_query”), effectively detecting mentions of the user’s query.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AFC

-0.07

베

-0.07

 ensemble

-0.06

-game

-0.06

 UserDao

-0.06

 Dalton

-0.06

DP

-0.06

WARE

-0.06

 pains

-0.06

/global

-0.06

POSITIVE LOGITS

属于

0.07

0.06

과의

0.06

 leggings

0.06

 izin

0.06

ใด

0.06

 كل

0.06

uku

0.06

ungi

0.06

 Sunder

0.06

Activations Density 0.039%

utter

The neuron activates on occurrences of the word “query” (and closely related tokens like “user_query”), effectively detecting mentions of the user’s query.

No Comments

No Known Activations