INDEX

Explanations

negations and expressions of disagreement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 myſelf

-1.05

Efq

-0.89

 uſed

-0.88

 itſelf

-0.87

 raiſ

-0.87

 ſeveral

-0.86

 Tacitus

-0.81

 Manchuria

-0.81

Portály

-0.80

 purpoſe

-0.79

POSITIVE LOGITS

is

0.99

not

0.91

Not

0.85

WAS

0.79

我不是

0.78

not

0.76

isn

0.73

不是

0.73

IsNot

0.71

 being

0.70

Activations Density 0.114%

negations and expressions of disagreement

No Comments

No Known Activations