INDEX

Explanations

mentions of "input" or variations thereof in a technical context

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

input methods and directories

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

 forState

-0.56

 zoude

-0.53

 للمعارف

-0.53

 Zeller

-0.51

azole

-0.50

ever

-0.50

brengen

-0.49

 queſta

-0.49

 mesmas

-0.48

assertIs

-0.48

POSITIVE LOGITS

 input

1.23

input

1.18

Input

1.15

 Input

1.13

INPUT

0.99

 INPUT

0.94

 getInput

0.91

 inputs

0.90

getInput

0.88

setInput

0.86

Activations Density 0.011%

mentions of "input" or variations thereof in a technical context

input methods and directories

No Comments

No Known Activations