INDEX

Explanations

specific qualifiers or emphasis terms related to particular conditions or contexts in discussions

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

especially

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

Ex

-0.39

Ex

-0.38

frag

-0.37

wh

-0.36

ex

-0.36

ex

-0.36

poli

-0.35

lables

-0.35

centa

-0.35

leh

-0.34

POSITIVE LOGITS



0.98

 protoimpl

0.83

цездатний

0.72

KommentareTeilen

0.65

 Normdatei

0.62

OGND

0.61

principalColumn

0.61

httphttps

0.61

 otomatig

0.59

новниш

0.58

Activations Density 0.070%

specific qualifiers or emphasis terms related to particular conditions or contexts in discussions

especially

No Comments

No Known Activations