INDEX

Explanations

Code/Log data

np_max-act · gemini-2.0-flash

The neuron detects speaker‐role tokens and the boundaries that mark the start of a user’s message (e.g. “user:” 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Enemy

-0.07

icates

-0.07

_contacts

-0.07

 ры

-0.06

 نیاز

-0.06

 правил

-0.06

 Nylon

-0.06

 тка

-0.06

 biraz

-0.06

 trot

-0.06

POSITIVE LOGITS

(deg

0.06

 resides

0.06

(fin

0.06

 цих

0.06

Copying

0.06

Perform

0.06

 العامة

0.06

XML

0.06

(open

0.06

 керів

0.06

Activations Density 0.010%

Code/Log data

The neuron detects speaker‐role tokens and the boundaries that mark the start of a user’s message (e.g. “user:” and related punctuation).

No Comments

No Known Activations