INDEX

Explanations

structured formats and headings within the text

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

headings and structural markup

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

 विश्वसनीयता

-0.58

CodeAttribute

-0.53

 للمعارف

-0.53

principalColumn

-0.50

 ErrInvalid

-0.50

TypedDataSet

-0.50

 deport

-0.50

Portail

-0.50

parsedMessage

-0.50

 <=",

-0.49

POSITIVE LOGITS

Hochspringen

0.40

发表于

0.35

 trúc

0.33

 κε

0.32

 hierarchy

0.32

<h1>

0.32

 volna

0.32

AutoScaleMode

0.31

édrale

0.31

 moim

0.31

Activations Density 0.009%

structured formats and headings within the text

headings and structural markup

No Comments

No Known Activations