INDEX

Explanations

references to questions and their responses in a structured format

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

Intent extra gets

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

tagext

-0.31

__

-0.28

-0.26

 Carpet

-0.26

 MonoBehaviour

-0.25

 musíte

-0.24

ClassNotFound

-0.24

 teren

-0.24

SerializedSize

-0.24

POSITIVE LOGITS

ValueStyle

0.68

<unused41>

0.66

<unused47>

0.66

<unused7>

0.66

<unused23>

0.66

<unused17>

0.66

<unused51>

0.66

<unused8>

0.66

<pad>

0.66

[@BOS@]

0.66

Activations Density 0.017%

references to questions and their responses in a structured format

Intent extra gets

No Comments

No Known Activations