INDEX

Explanations

points of data or code-related structures that signify specific commands or functionalities

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

code block delimiters

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

 Administrativna

-0.88

ValueStyle

-0.65

delwed

-0.65

 חיצוניים

-0.61

PYX

-0.60

 يتيمه

-0.60

 <>",

-0.56

PyExc

-0.55

цездатний

-0.55

PreferredItem

-0.54

POSITIVE LOGITS

IntoConstraints

0.48

ſelf

0.41

 Inſ

0.40

 ſtill

0.39

intenant

0.38

 laſt

0.35

 Chriftian

0.35

 Reſ

0.35

 ſtand

0.35

 preſent

0.35

Activations Density 0.015%

points of data or code-related structures that signify specific commands or functionalities

code block delimiters

No Comments

No Known Activations