INDEX

Explanations

phrases related to ideas and information retrieval

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

finding new information

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

 online

-0.35

 prior

-0.33

 bale

-0.33

 Insertion

-0.32

tamine

-0.32

LCA

-0.32

 getattr

-0.32

lou

-0.31

ropolitan

-0.31

 mens

-0.31

POSITIVE LOGITS

 незавершена

0.86

RegressionTest

0.71

 valuable

0.59

 useful

0.57

useful

0.57

Useful

0.53

 discoveries

0.52

ftagPool

0.52

 revelations

0.52

valuable

0.51

Activations Density 0.091%

phrases related to ideas and information retrieval

finding new information

No Comments

No Known Activations