INDEX

Explanations

phrases related to accusations and allegations of wrongdoing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

+#+#

-0.81

:+:

-0.74

AndEndTag

-0.68

Біографія

-0.66

談社

-0.65



-0.64

évaluateur

-0.62

atron

-0.61

httphttps

-0.61

ilever

-0.61

POSITIVE LOGITS

 accused

1.19

 accuse

1.06

 accuses

1.05

 accusing

1.00

 accusations

0.94

 acusa

0.84

 accusation

0.83

 acusado

0.81

 CreateTagHelper

0.80

accus

0.79

Activations Density 0.256%

phrases related to accusations and allegations of wrongdoing

No Comments

No Known Activations