INDEX

Explanations

negative prefixes and their associated constructs

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

un- or de- prefixes

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 autorytatywna

-0.93

iſen

-0.92

ſicht

-0.91

ſchaft

-0.91

 dezelve

-0.89

<unused43>

-0.89

parsedMessage

-0.89

<unused74>

-0.89

niſſe

-0.89

[@BOS@]

-0.88

POSITIVE LOGITS

re

0.43

↵

0.37

“

0.35

<eos>

0.34

0.33

↵↵

0.33

re

0.33

Re

0.32

Activations Density 0.007%

negative prefixes and their associated constructs

un- or de- prefixes

No Comments

No Known Activations