INDEX

Explanations

you followed by auxiliary

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

references to people via personal pronouns, with a strong preference for second-person address (you/your) and weaker responses to first- and third-person forms.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.3-70B-IT @ 50-resid-post-gf

Configuration

Goodfire/Llama-3.3-70B-Instruct-SAE-l50/Llama-3.3-70B-Instruct-SAE-l50.pt

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ought

-0.09

gether

-0.08

ned

-0.08

mai

-0.08

itra

-0.08

culo

-0.08

 emblem

-0.08

Ø¯Ø§Ø±

-0.08

aset

-0.08

able

-0.07

POSITIVE LOGITS

åĢĳ

0.17

zelf

0.16

’;re

0.13

self

0.13

're

0.13

-mÃªme

0.12

've

0.12

’;ve

0.12

’;ll

0.12

'll

0.11

Activations Density 0.316%

you followed by auxiliary

references to people via personal pronouns, with a strong preference for second-person address (you/your) and weaker responses to first- and third-person forms.

No Comments

No Known Activations

you followed by auxiliary

references to people via personal pronouns, with a strong preference for second-person address (you/your) and weaker responses to first- and third-person forms.

No Comments

No Known Activations