INDEX

Explanations

negative assessments of groups, particularly in political contexts

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

just another

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.20.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

 defaultstate

-0.89

 Италијани

-0.85

TypedDataSet

-0.84

 nakalista

-0.83

featureID

-0.80

IsContent

-0.78

PYX

-0.78

Rujuakan

-0.77

WriteBarrier

-0.77

CodedInputStream

-0.74

POSITIVE LOGITS

 another

3.12

another

2.81

Another

2.64

 Another

2.61

 ANOTHER

2.27

 otra

1.74

 otro

1.72

Otro

1.62

Otra

1.61

 Otro

1.58

Activations Density 0.211%

negative assessments of groups, particularly in political contexts

just another

No Comments

No Known Activations