INDEX

Explanations

references to religious or spiritual texts and their interpretations

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

constructs and references

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

 oprot

-0.81

 незавершена

-0.75

 ویکی‌پدیا

-0.71

Personendaten

-0.64

⟬

-0.60

الحياه

-0.56

AsUp

-0.55

Geplaatst

-0.54

exitRule

-0.52

 kasarigan

-0.51

POSITIVE LOGITS

 referencing

0.34

tagext

0.34

 random

0.33

transQ

0.33

 references

0.33

References

0.32

 miss

0.32

 stitches

0.31

 batteries

0.31

 refer

0.31

Activations Density 0.082%

references to religious or spiritual texts and their interpretations

constructs and references

No Comments

No Known Activations