INDEX

Explanations

negations and expressions of inability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

 Want

-0.43

 wanting

-0.43

 incident

-0.41

'{@

-0.41

 accident

-0.40

 needing

-0.39

 Paglinawan

-0.38

try

-0.37

 encounters

-0.37

 intentar

-0.37

POSITIVE LOGITS

 afford

0.96

 tolerate

0.87

afford

0.81

 withstand

0.75

 Afford

0.68

toler

0.68

 tolerated

0.66

 reliably

0.66

 toler

0.65

 handle

0.65

Activations Density 0.359%

negations and expressions of inability

No Comments

No Known Activations