INDEX

Explanations

the

np_max-act · gemini-2.0-flash

The neuron weakly activates on numeric and symbol‐heavy tokens (e.g. years, measurements, unit symbols, digits and related punctuation).

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 intents

-0.08

Sys

-0.07

 Παν

-0.07

_NM

-0.06

 wastes

-0.06

なが

-0.06

 spol

-0.06

Tor

-0.06

居

-0.06

 Compact

-0.06

POSITIVE LOGITS

(ctx

0.07

числ

0.06

.controllers

0.06

 Presented

0.06

①

0.06

stab

0.06

 رأ

0.06

 Jays

0.06

něji

0.06

 BadRequest

0.06

Activations Density 0.000%

the

The neuron weakly activates on numeric and symbol‐heavy tokens (e.g. years, measurements, unit symbols, digits and related punctuation).

No Comments

No Known Activations