Skip to content

Rewrite new simulator to use JSON mode; additional fixes to new simulator #34

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
hijohnnylin wants to merge 14 commits into
base: main
Choose a base branch
from
Open

Rewrite new simulator to use JSON mode; additional fixes to new simulator #34

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
8d260b3
handle 3 edge cases for new simulator
hijohnnylin Dec 29, 2023
67e4f99
Allow float simulated activation values and handle non-numeric values
hijohnnylin Dec 29, 2023
208663e
Enforce predicted activation float range
hijohnnylin Dec 29, 2023
9a6739d
Allow approximate space matching with all tokens, not just the first one
hijohnnylin Dec 29, 2023
d2b5ef5
Fix bug with last activation values
hijohnnylin Dec 30, 2023
4ec7d74
Further fix edge case 3 to set correct start_line_index when first to…
hijohnnylin Dec 30, 2023
610fdce
Add debug logging
hijohnnylin Dec 30, 2023
9759a15
Better catching of non-float activation values
hijohnnylin Dec 31, 2023
66dfe51
Catch cases where GPT returns a token line with an invalid format
hijohnnylin Dec 31, 2023
c7c6eea
Rewrite new scorer with JSON mode. Add temperature = 0.
hijohnnylin Jan 5, 2024
62eedb5
Fix: Catch incorrect type of activation value in JSON response
hijohnnylin Jan 6, 2024
58af073
Increase max_tokens to 2000, move endoftext and non-ascii handling ou…
hijohnnylin Jan 8, 2024
d62da4e
more detailed error logging for easier debugging of invalid GPT respo…
hijohnnylin Jan 10, 2024
bfbeb57
ENHANCEMENT: Fine-tuned scorer: Update prompt, add new examples
hijohnnylin Feb 14, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion neuron-explainer/neuron_explainer/api_client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,7 +114,7 @@ def __init__(

@exponential_backoff(retry_on=is_api_error)
async def make_request(
self, timeout_seconds: Optional[int] = None, **kwargs: Any
self, timeout_seconds: Optional[int] = None, json_mode: Optional[bool] = False, **kwargs: Any
) -> dict[str, Any]:
if self._cache is not None:
key = orjson.dumps(kwargs)
Expand All @@ -130,6 +130,8 @@ async def make_request(
# endpoint. Otherwise, it should be sent to the /completions endpoint.
url = BASE_API_URL + ("/chat/completions" if "messages" in kwargs else "/completions")
kwargs["model"] = self.model_name
if json_mode:
kwargs["response_format"] = {"type": "json_object"}
response = await http_client.post(url, headers=API_HTTP_HEADERS, json=kwargs)
# The response json has useful information but the exception doesn't include it, so print it
# out then reraise.
Expand Down
179 changes: 179 additions & 0 deletions neuron-explainer/neuron_explainer/explanations/few_shot_examples.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,7 @@ class FewShotExampleSet(Enum):
ORIGINAL = "original"
NEWER = "newer"
TEST = "test"
JL_FINE_TUNED = "jl_fine_tuned"

@classmethod
def from_string(cls, string: str) -> FewShotExampleSet:
Expand All @@ -56,6 +57,8 @@ def get_examples(self) -> list[Example]:
return NEWER_EXAMPLES
elif self is FewShotExampleSet.TEST:
return TEST_EXAMPLES
elif self is FewShotExampleSet.JL_FINE_TUNED:
return JL_FINE_TUNED_EXAMPLES
else:
raise ValueError(f"Unhandled example set: {self}")

Expand Down Expand Up @@ -1038,3 +1041,179 @@ def get_single_token_prediction_example(self) -> Example:
token_index_to_score=18,
explanation="instances of the token 'ate' as part of another word",
)


JL_FINE_TUNED_EXAMPLES = [
Example(
activation_records=[
ActivationRecord(
tokens=[
"The",
" cat",
" jumped",
" on",
" my",
" laptop",
".",
],
activations=[
0,
0,
0,
0,
0,
0,
0
],
),
],
first_revealed_activation_indices=[],
explanation="the word \"laptop\" before the word \"cat\"",
),
Example(
activation_records=[
ActivationRecord(
tokens=[
"The",
" cat",
" jumped",
" on",
" my",
" laptop",
".",
],
activations=[
0,
10,
0,
0,
0,
0,
0
],
),
],
first_revealed_activation_indices=[],
explanation="the word \"cat\" before the word \"laptop\"",
),
Example(
activation_records=[
ActivationRecord(
tokens=[
"I",
" am",
" using",
" a",
" keyboard",
".",
],
activations=[
0,
0,
0,
0,
10,
0
],
),
],
first_revealed_activation_indices=[],
explanation="the word before a period",
),
Example(
activation_records=[
ActivationRecord(
tokens=[
"The",
" sun",
" is",
" shining",
".",
" The",
" clouds",
" are",
" gone",
".",
" Great",
" weather",
"!",
],
activations=[
0,
0,
0,
10,
0,
0,
0,
0,
10,
0,
0,
0,
0
],
),
],
first_revealed_activation_indices=[],
explanation="the word before period",
),
]

NEWER_SINGLE_TOKEN_EXAMPLE = Example(
activation_records=[
ActivationRecord(
tokens=[
"B",
"10",
" ",
"111",
" MON",
"DAY",
",",
" F",
"EB",
"RU",
"ARY",
" ",
"11",
",",
" ",
"201",
"9",
" DON",
"ATE",
"fake higher scoring token", # See below.
],
activations=[
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0.37,
# This fake activation makes the previous token's activation normalize to 8, which
# might help address overconfidence in "10" activations for the one-token-at-a-time
# scoring prompt. This value and the associated token don't actually appear anywhere
# in the prompt.
0.45,
],
),
],
first_revealed_activation_indices=[],
token_index_to_score=18,
explanation="instances of the token 'ate' as part of another word",
)
Loading