INDEX

Explanations

co-occurrences of the word "and" in various contexts

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

men and women

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

もあった

-0.41

 estadia

-0.39

 pelaksanaan

-0.38

もある

-0.38

 control

-0.37

 arranque

-0.37

 Höhe

-0.37

 Gründ

-0.36

 aislada

-0.36

 estadía

-0.35

POSITIVE LOGITS

ImageContext

0.80

tanleria

0.63

 persons

0.63

RegressionTest

0.63

parsedMessage

0.62

ſelves

0.61

EDEFAULT

0.60

AccessorTable

0.59

 perfons

0.59

Personensuche

0.58

Activations Density 0.086%

co-occurrences of the word "and" in various contexts

men and women

No Comments

No Known Activations