INDEX

Explanations

highly technical and formal language related to mathematical proofs and statements

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

note that

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

 initComponents

-0.82

GOTREF

-0.81

 يتيمه

-0.74

IntoConstraints

-0.73

ReusableCell

-0.72

 للمعارف

-0.72

DockStyle

-0.71

<unused41>

-0.69

<unused23>

-0.69

[@BOS@]

-0.69

POSITIVE LOGITS

<eos>

0.40

↵↵

0.31

 Walkover

0.30

 ervoor

0.30

 Part

0.29

Gruss

0.29

 lateinit

0.29

0.28

qJ

0.28

The

0.27

Activations Density 0.003%

highly technical and formal language related to mathematical proofs and statements

note that

No Comments

No Known Activations