INDEX

Explanations

tragedies

np_max-act · gemini-2.0-flash

This neuron detects the special chat‐format control tokens (like <|start_header_id|>, <|end_header_id|>, <|begin_of_text|>, and <|eot_id|>) that mark message or header boundaries.

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

 friday

-0.07

 hard

-0.06

 hiểu

-0.06

 ruling

-0.06

 connected

-0.06

еси

-0.06

 thuế

-0.06

orrent

-0.06

 Bender

-0.06

简单

-0.06

POSITIVE LOGITS

 گر

0.07

mun

0.07

 Async

0.06

(sf

0.06

.setView

0.06

ังกล

0.06

�

0.06

 presets

0.06

 derog

0.06

 demonstr

0.06

Activations Density 0.037%

tragedies

This neuron detects the special chat‐format control tokens (like <|start_header_id|>, <|end_header_id|>, <|begin_of_text|>, and <|eot_id|>) that mark message or header boundaries.

No Comments

No Known Activations

tragedies

This neuron detects the special chat‐format control tokens (like <|start_header_id|>, <|end_header_id|>, <|begin_of_text|>, and <|eot_id|>) that mark message or header boundaries.

No Comments

No Known Activations