INDEX

Explanations

references to actions associated with caring or helping others in crisis situations

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

emotional reactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.16.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

AutoresizingMask

-0.63

rind

-0.62

いきましょう

-0.60

出版年

-0.59

ThroughAttribute

-0.59

 undesirable

-0.59

resave

-0.57

AILABILITY

-0.57

 discard

-0.57

 experimentation

-0.56

POSITIVE LOGITS

 thanking

0.76

 heartwarming

0.72

 thanked

0.69

 hugged

0.66

感動

0.66

 sobbed

0.62

 gratitude

0.62

 touched

0.61

 THANK

0.60

 comforted

0.59

Activations Density 0.126%

references to actions associated with caring or helping others in crisis situations

emotional reactions

No Comments

No Known Activations