INDEX

Explanations

Code

np_max-act · gemini-2.0-flash

The neuron fires on programming code snippets, i.e. tokens characteristic of source‐code (identifiers, keywords, punctuation) rather than plain text.

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

 potvr

-0.08

τευ

-0.07

Storyboard

-0.06

 strangely

-0.06

色

-0.06

yscale

-0.06

 місцев

-0.06

 jewelry

-0.06

 unanswered

-0.06

estruct

-0.06

POSITIVE LOGITS

_API

0.07

 yapı

0.06

>(↵

0.06

(cost

0.06

<S

0.06

<Long

0.06

.XRLabel

0.06

\Entities

0.06

Brand

0.06

Activations Density 0.030%

Code

The neuron fires on programming code snippets, i.e. tokens characteristic of source‐code (identifiers, keywords, punctuation) rather than plain text.

No Comments

No Known Activations