INDEX

Explanations

expressions of agreement or consensus

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

statements of agreement

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

 Petersen

-0.51

 Thors

-0.50

нент

-0.49

ベーション

-0.46

 Garuda

-0.45

vh

-0.45

parentElement

-0.44

-0.43

POSITIVE LOGITS

 agree

1.04

agree

1.02

 Agree

0.96

 agrees

0.94

Agree

0.93

 agreeing

0.88

 agreed

0.87

agreed

0.86

 Agreed

0.84

Disagree

0.73

Activations Density 0.017%

expressions of agreement or consensus

statements of agreement

No Comments

No Known Activations