INDEX

Explanations

code constructs related to object or class definitions

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

first occurrences

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

RenderAtEndOf

-0.86

didReceive

-0.84

<unused41>

-0.77

<unused17>

-0.77

<unused42>

-0.76

<unused16>

-0.76

<unused51>

-0.76

<unused52>

-0.76

[@BOS@]

-0.76

<unused14>

-0.76

POSITIVE LOGITS

 Treiber

0.30

 Bielefeld

0.30

 Arteta

0.30

первых

0.29

ISTRATION

0.29

 Utara

0.27

0.26

Activations Density 0.065%

code constructs related to object or class definitions

first occurrences

No Comments

No Known Activations