INDEX

Explanations

Language Models

np_max-act · gemini-2.0-flash

The neuron fires on occurrences of specific transformer‐based model names (e.g. BERT, RoBERTa, XLNet, ELECTRA).

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

Suffix

-0.06

'>$

-0.06

-quote

-0.06

редит

-0.06

 retorna

-0.06

 berth

-0.06

 kadar

-0.06

abant

-0.06

 yönelik

-0.06

Donald

-0.06

POSITIVE LOGITS

 nowrap

0.06

 vines

0.06

 inventive

0.06

 moot

0.06

臺

0.06

 برنامه

0.06

 prize

0.06

 --------------------------------------------------------------------------↵

0.06

EW

0.06

 fencing

0.06

Activations Density 0.002%

Language Models

The neuron fires on occurrences of specific transformer‐based model names (e.g. BERT, RoBERTa, XLNet, ELECTRA).

No Comments

No Known Activations