4 ms·

Are they charging for the guardrails? Like do the guardrails expend token counts to then block you from the output of other tokens?

SOLAR_FIELDS 5mo ago
Not directly, as it comes in as a not charged error but the weighted generation path used until you hit the guardrail is basically wasted tokens, so yes, indirectly. If I hit a guardrail and rewind I’ve found the training will still be biased towards guardrailing out if you rewind one turn. Rewinding multiple turns allows steering away from that path, but all of the original token spend down that path is wasted
kay_o 5mo ago
When your session is force ended for "abuse" you get neither the response nor a refund Security, games (think weapons, PVP, attacking, etc), sometimes even asking it for a security review of some CRUD code it wrote itself
- danpalmer 5mo ago
  What a joke. Must make it pretty easy to poison a session, you don't need to persuade the model about anything, just trigger its security controls, ideally after as much context as possible, but before it has generated any useful output.
  - kay_o 5mo ago
    After all, what is roleplay or games but a jailbreak of guard rails? :] I've even had it refuse CTFs knowing it is a CTF with blatantly obvious CTF flag, no actual application
- bombcar 5mo ago
  I asked it about a “yellow background cell” in Excel and it spewed a book at me. Then it solved the issue.
acters 5mo ago
Yes tokens used (input and sometimes output) are always charged. You likely get charged for the preloaded system prompt, too.
jerrythegerbil 5mo ago
Yes. When certain keywords are matched or topics, there is a warning transparently injected server side appended to the system prompt of the convo that’s miles long. It is injected and reevaluated every tool call. If you begin a generic reverse engineering task, 30+ tool calls in a row. The moment it sees something it doesn’t like, token burn, single tool calls iteration, “This is a known CTF challenge, I can proceed”, single tool calls iteration, “This is a real CTF challenge, I can proceed”, etc. It’s heavily neutered now, without changing the model, and you pay for the privilege and don’t notice. The end result of course being that it both expensive and useless for approved CTF tasks. No one is using Opus for security. If they think it’s working, the harsh reality is they’re not doing security work; they’re just generically finding bugs. I do this for a job and can demonstrate this plain as day, dump the injected prompt, and notice what it’s doing isn’t security work, it just looks like it. Happy to write a blog about it if you want to know more. Apparently many people think it’s working for them when it absolutely isn’t.
- Khaine 5mo ago
  I would find a blog post on this really interesting.
- bombcar 5mo ago
  Mythos turns out to be Opus 4.8 in a trenchcoat with guardrails removed.
  - satvikpendem 5mo ago
    Opus 4.7 and 4.8 are well known to be distilled versions of Mythos unlike 4.6 which is why they are rated so badly by users compared to 4.6.
- ramblin_prose 5mo ago
  I'd like to read that blog please! Thanks for the insight.
gmerc 5mo ago
Of course they are. It's standard SaaS to charge for security features ;)

SOLAR_FIELDS 5mo ago

Not directly, as it comes in as a not charged error but the weighted generation path used until you hit the guardrail is basically wasted tokens, so yes, indirectly. If I hit a guardrail and rewind I’ve found the training will still be biased towards guardrailing out if you rewind one turn. Rewinding multiple turns allows steering away from that path, but all of the original token spend down that path is wasted

kay_o 5mo ago

When your session is force ended for "abuse" you get neither the response nor a refund Security, games (think weapons, PVP, attacking, etc), sometimes even asking it for a security review of some CRUD code it wrote itself

danpalmer 5mo ago
What a joke. Must make it pretty easy to poison a session, you don't need to persuade the model about anything, just trigger its security controls, ideally after as much context as possible, but before it has generated any useful output.
- kay_o 5mo ago
  After all, what is roleplay or games but a jailbreak of guard rails? :] I've even had it refuse CTFs knowing it is a CTF with blatantly obvious CTF flag, no actual application
bombcar 5mo ago
I asked it about a “yellow background cell” in Excel and it spewed a book at me. Then it solved the issue.

acters 5mo ago

Yes tokens used (input and sometimes output) are always charged. You likely get charged for the preloaded system prompt, too.

jerrythegerbil 5mo ago

Yes. When certain keywords are matched or topics, there is a warning transparently injected server side appended to the system prompt of the convo that’s miles long. It is injected and reevaluated every tool call. If you begin a generic reverse engineering task, 30+ tool calls in a row. The moment it sees something it doesn’t like, token burn, single tool calls iteration, “This is a known CTF challenge, I can proceed”, single tool calls iteration, “This is a real CTF challenge, I can proceed”, etc. It’s heavily neutered now, without changing the model, and you pay for the privilege and don’t notice. The end result of course being that it both expensive and useless for approved CTF tasks. No one is using Opus for security. If they think it’s working, the harsh reality is they’re not doing security work; they’re just generically finding bugs. I do this for a job and can demonstrate this plain as day, dump the injected prompt, and notice what it’s doing isn’t security work, it just looks like it. Happy to write a blog about it if you want to know more. Apparently many people think it’s working for them when it absolutely isn’t.

Khaine 5mo ago
I would find a blog post on this really interesting.
bombcar 5mo ago
Mythos turns out to be Opus 4.8 in a trenchcoat with guardrails removed.
- satvikpendem 5mo ago
  Opus 4.7 and 4.8 are well known to be distilled versions of Mythos unlike 4.6 which is why they are rated so badly by users compared to 4.6.
ramblin_prose 5mo ago
I'd like to read that blog please! Thanks for the insight.

gmerc 5mo ago

Of course they are. It's standard SaaS to charge for security features ;)