INDEX

Explanations

Punctuation

np_max-act · gemini-2.0-flash

instances of uncertainty or requests for clarification.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

This neuron flags the special conversation‐structure tokens marking the end of a user or assistant turn (e.g. the “<|eot_id|>” end‐of‐turn marker and related header delimiters).

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

max

-0.06

.contains

-0.06

 fused

-0.06

 titles

-0.06

短

-0.06

 بها

-0.06

ENO

-0.06

 discouraged

-0.06

enght

-0.06

 mixer

-0.06

POSITIVE LOGITS

_about

0.07

"><!--

0.06

mesine

0.06

）

0.06

》（

0.06

.GUI

0.06

.ul

0.06

!',↵

0.06

?>"/>↵

0.06

 inşa

0.06

Activations Density 0.045%

Punctuation

instances of uncertainty or requests for clarification.

This neuron flags the special conversation‐structure tokens marking the end of a user or assistant turn (e.g. the “<|eot_id|>” end‐of‐turn marker and related header delimiters).

No Comments

No Known Activations

Punctuation

instances of uncertainty or requests for clarification.

This neuron flags the special conversation‐structure tokens marking the end of a user or assistant turn (e.g. the “<|eot_id|>” end‐of‐turn marker and related header delimiters).

No Comments

No Known Activations