INDEX

Explanations

references to file operations and content-related metadata

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

image descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

ck

-0.38

Scale

-0.34

<eos>

-0.34

scale

-0.32

tz

-0.32

만

-0.32

USE

-0.31

tw

-0.31

loc

-0.31

smarty

-0.31

POSITIVE LOGITS

expandindo

0.76

UserScript

0.72

 kasarigan

0.71

 beginnetje

0.67

IUrlHelper

0.64

+#+#

0.63

@"/

0.63

 autorytatywna

0.62

 iprot

0.61

Diweddarwch

0.61

Activations Density 0.004%

references to file operations and content-related metadata

image descriptions

No Comments

No Known Activations