INDEX

Explanations

various testing or error checking terms in code-related contexts

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

test and ref markers

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

 Савезне

-0.43

 Grüsse

-0.40

 ſever

-0.36

 JpaRepository

-0.35

 themſelves

-0.31

 RIPRODUZIONE

-0.31

colgante

-0.30

ngdoc

-0.30

engaruhi

-0.29

 commandement

-0.29

POSITIVE LOGITS

Cob

0.60

maco

0.59



0.58

Logic

0.57

<unused41>

0.57

<unused3>

0.57

 cyber

0.57

 Cyril

0.57

<pad>

0.57

[@BOS@]

0.57

Activations Density 0.071%

various testing or error checking terms in code-related contexts

test and ref markers

No Comments

No Known Activations