INDEX

Explanations

terms related to input data and parameters

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

input fields or data

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

hood

-0.44

 AssemblyTitle

-0.44

 épaules

-0.42

GrantedAuthority

-0.42

舺

-0.40

thâu

-0.40

TestingModule

-0.40

contentLoaded

-0.39

 pthread

-0.38

tahankan

-0.38

POSITIVE LOGITS

 input

1.06

 inputs

0.90

 Input

0.84

 getInput

0.84

input

0.81

getInput

0.76

 Inputs

0.75

 INPUT

0.71

INPUT

0.70

Input

0.69

Activations Density 0.039%

terms related to input data and parameters

input fields or data

No Comments

No Known Activations