INDEX

Explanations

Summarization

np_max-act · gemini-2.0-flash

This neuron detects the discourse markers that introduce questions or summaries – e.g. interrogative words (“What,” “are,” “major,” “points”) in a user query and the typical “The article…”/“Summary” tokens at the start of an assistant’s summary.

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

-0.07

.do

-0.07

 Tribe

-0.06

 exhaustion

-0.06

method

-0.06

fot

-0.06

dan

-0.06

pit

-0.06

jectories

-0.06

 един

-0.06

POSITIVE LOGITS

�

0.07

_pwd

0.06

regon

0.06

.jdesktop

0.06

เพล

0.06

�

0.06

 Maur

0.06

ισε

0.06

隔

0.06

decrypt

0.06

Activations Density 0.061%

Summarization

This neuron detects the discourse markers that introduce questions or summaries – e.g. interrogative words (“What,” “are,” “major,” “points”) in a user query and the typical “The article…”/“Summary” tokens at the start of an assistant’s summary.

No Comments

No Known Activations

Summarization

This neuron detects the discourse markers that introduce questions or summaries – e.g. interrogative words (“What,” “are,” “major,” “points”) in a user query and the typical “The article…”/“Summary” tokens at the start of an assistant’s summary.

No Comments

No Known Activations