INDEX

Explanations

references to self-harm or suicidal ideation

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ſelf

-0.45

 themſelves

-0.42

 natives

-0.36

felves

-0.36

TRAIT

-0.36

 للمعارف

-0.35

ſelves

-0.35

 himſelf

-0.35

 ſtand

-0.35

 Ubicación

-0.34

POSITIVE LOGITS

__*/

0.49

 GetEnumerator

0.45

قایناق‌لار

0.44

DataContract

0.44

 COMMIT

0.42

zarchiwizowane

0.42

いしい

0.42

しゃれ

0.41

 CreateTagHelper

0.40

Referencer

0.40

Activations Density 0.084%

references to self-harm or suicidal ideation

No Comments

No Known Activations