INDEX

Explanations

punctuation marks in proximity to text

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

semicolons and commas

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

MF

-0.35

发表于

-0.32

-0.31

-0.28

Base

-0.28

-0.27

 viñ

-0.27

France

-0.26

The

-0.26

 mérite

-0.26

POSITIVE LOGITS

ftagPool

0.88

 كومونز

0.77

ſehen

0.75

niſſe

0.73

ſicht

0.73

⟬

0.72

 ſei

0.72

enablog

0.72

linawan

0.72

iſche

0.71

Activations Density 0.037%

punctuation marks in proximity to text

semicolons and commas

No Comments

No Known Activations