INDEX

Explanations

references to specific dates and celebrations, particularly in relation to January and February

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

January and February dates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 Sept

-0.45

中秋

-0.42

zeptember

-0.42

 SEPTEMBER

-0.42

SEPTEMBER

-0.42

 September

-0.41

 July

-0.40

 augusztus

-0.40

 wrze

-0.40

 september

-0.39

POSITIVE LOGITS

February

0.90

 February

0.88

Feb

0.85

Feb

0.82

January

0.79

 January

0.78

FEB

0.77

FEBRUARY

0.77

 FEBRUARY

0.76

feb

0.73

Activations Density 0.121%

references to specific dates and celebrations, particularly in relation to January and February

January and February dates

No Comments

No Known Activations