INDEX

Explanations

sequences of code snippets with varied import statements and configurations

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

programming imports

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

'{@

-0.89

featureID

-0.80

+:+

-0.78

EDEFAULT

-0.78

mpagne

-0.76

AnimationsModule

-0.75

 snippetHide

-0.75

UnusedPrivate

-0.74

تقاوى

-0.74

setVerticalGroup

-0.74

POSITIVE LOGITS

 vandens

0.36

jį

0.36

 juos

0.35

volumen

0.32

 auroit

0.32

 Usuarios

0.31

ją

0.31

 fidèles

0.31

 Inscrivez

0.30

</sub>

0.30

Activations Density 0.004%

sequences of code snippets with varied import statements and configurations

programming imports

No Comments

No Known Activations