INDEX

Explanations

categories and classifications within a structured context

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

Category:

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

 zoude

-0.49

elemField

-0.47

 ControllerBase

-0.46

 awakeFromNib

-0.45

⟬

-0.45

RenderAtEndOf

-0.44

:+:

-0.44

 msglen

-0.43

şört

-0.43

 aDecoder

-0.43

POSITIVE LOGITS



0.50

Autoritní

0.43

bkz

0.41

Хьажоргаш

0.39

انجليز

0.38

underbrace

0.37

Efq

0.36

ğraf

0.36

Gé

0.36

 Photocase

0.36

Activations Density 0.008%

categories and classifications within a structured context

Category:

No Comments

No Known Activations