INDEX

Explanations

phrases related to summarizing and presenting research findings

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

information presented in figures or tables

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

 anything

-0.39

Luc

-0.39

 deco

-0.38

 frapp

-0.38

 placenta

-0.36

исленность

-0.36

GenerationType

-0.36

RegressionTest

-0.36

 laguna

-0.36

fal

-0.35

POSITIVE LOGITS

 depicted

0.66

 shown

0.65

 illustrated

0.64

shown

0.62

 outlined

0.62

 pictured

0.60

illustrated

0.59

 summarised

0.57

devamını

0.57

 illustratie

0.57

Activations Density 0.219%

phrases related to summarizing and presenting research findings

information presented in figures or tables

No Comments

No Known Activations