INDEX

Explanations

phrases indicating general consensus or widely accepted ideas

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

generally accepted facts

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

intéress

-0.40

 architet

-0.38

 isolé

-0.36

erapeu

-0.35

therapeutic

-0.35

 Patienten

-0.35

 purpoſe

-0.35

 cauſe

-0.34

berday

-0.33

 déta

-0.33

POSITIVE LOGITS

 consensus

0.66

 accepted

0.65

 prevailing

0.61

 Consensus

0.59

 считается

0.57

consensus

0.57

accepted

0.55

 принято

0.54

GEBURTSDATUM

0.53

ACCEPTED

0.52

Activations Density 0.068%

phrases indicating general consensus or widely accepted ideas

generally accepted facts

No Comments

No Known Activations